Skip to main content

Full text of "Genetic analysis of intraspecific variation in pathovars of Xanthomonas campestris"

See other formats


GENETIC ANALYSIS OF INTRASPECIFIC VARIATION IN 
PATHOVARS OF Xant±iC8nonas campestris 



BY 
GERARD RAYMOtO LAZO 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 
OF THE UNIVERSITY OF FLORIDA 
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS 
■ FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 
1987 



Copyright 1987 

by 

Gerard Raymond Lazo 



ACfCNOWLEDGMENTS 

I wish to express my gratitude to Dr. D. W. Gabriel, chairman of 
the advisory corrmittee, for his friendship, help and encouragement 
during the course of this research. Thanks are also expressed to all 
personnel associated with the laboratory, who were very generous and 
cooperative in supporting the research efforts which made this 
presentation possible. Special thanks are extended to A. Williams who 
assisted with the race-specificity studies on cotton; G. V. Minsavage 
who assisted with the SDS-PAGE analysis; R. Roffey, who assisted with 
the RFLP analysis; and A. Burges who assisted with various other 
aspects of this study. Thanks are also expressed to Drs. R. E. Stall, 
D. R. Pring, and M. J. Bassett for serving on the advisory committee 
and for their support. Appreciation is also acknowledged for 
financial support from the Oklahoma and Florida Agricultural 
Experiment Stations, and United States Department of Agriculture grant 
USDA-58-7B30- 3-465. And finally, I am grateful to my loving wife, 
Maria, for supporting and encouraging me throughout this degree 
program, for she tolerated my late hours and comforted me in the most 
difficult of times. 



Ill 



TABLE OF CCNTENTS 

Page 

ACKNOWLEDGMENTS m 

LIST OF TABLES vi 

LIST OF FIOJRES vii 

ABSTRACT ix 

CHAPTER 

ONE INTRODUCTION ]_ 

TWO LITERATURE REVIEW 4 

THREE HOST RANGE OF PATHOVARS OF XANTHOMONAS CAMPESTRIS 9 

Intrcducti on , ^ , 9 

Materials and Methods 10 

Bacterial Strains and Host Plants 10 

Plant Inoculations 12 

Physiological Differentiation 12 

Polyacrylamide Gel Electrophoresis 13 

Results 13 

Pathogenicity Tests 13 

Physiological Differentiation 20 

Polyacrylamide Gel Electrophoresis 22 

Discussion 22 

FOUR CONSERVATION OF PLASMID DNA SEQUENCES AND PATHOVAR 

IDENTIFICATION OF STRAINS OF XANTHOMONAS CAMPESTRIS 27 

Introducti on 27 

Materials and Methods 28 

Bacterial Strains 28 

Plasmid Extraction and Visualization 31 

Cloning of Plasmid 

Restriction Endonuclease Fragments 32 

DNA/DNA Hybridization 32 

Results 33 

Detection of Plasmid DNA 33 

Restriction Endonuclease Profiles 35 

DNA/DNA Hybridization 40 

Dot-blot Hybridization 42 

Discussi on 45 



IV 



FIVE ARE RACE-SPECIFIC AVIRULENCE GENES ENCODED BY PLASMIDS 

OF XANTH0MONA5 CAMPESTRIS FV. MALVACEARUM ? 50 

Introducti on 50 

Materials and Methods 51 

Bacterial Strains and Host Plants 51 

Plant Inoculations 51 

Plasmid Analysis 5I 

Plasraid Curing 53 

Race-Specif 3 city Genes 53 

Plasmid Origin of Replication and Mobilization 54 

Results 55 

Plasmids and Race-Specificity 55 

Plasmid Curing 58 

Race-Specif icity Genes 58 

Plasmid Origin of Replication and Mobilization 61 

Discussion g5 

SIX PATHOVARS OF XANTHOMONAS CAMPESTRIS ARE DISTINGUISHABLE 

BY RESTRICTION FRAGMENT LENGTH POLYMORPHISM 70 

Introduction 70 

Materials and Methods 72 

Bacterial Strains 72 

DNA Extracti on 72 

Agarose Gel Electrophoresis 75 

DNA Probes 75 

DNA Hybridization 76 

Restriction Fragment Patterns and Densitometry .... 77 

Results „ ^ ^ ^ 77 

Agarose Gel Electrophoresis 77 

DNA Probes 80 

DNA Hybridization „ 80 

Discussion , 88 

SEVEN SUMMARY 92 

Intraspecific Variation 92 

Plasmids 92 

Restriction Fragment Length Polymorphism 93 

Conclusions 95 



LITERATURE CITED 



97 

BIOGRAPHICAL SKETCH 106 



V 



LIST OF TABLES 
Table 



Paqe 



3-1 Strains of X. campestris used in host range investigation ... 11 

3-2 Legume plant reactions to inoculation with pathovars 

of X. campestris 2.4 

3-3 Host range observed for pathovars of X. campestris 

pathogenic to legume host plants ... .7. .7777777777 17 

3-4 Malvaceous host plant reactions to inoculation 

with strains of X. campestris pv. malvacearum 19 

3-5 Physiological reactions of strains of X. campestris 21 

4-1 Strains of X. campestris used for plasmid analysis 29 

4-2 Detection of plasmid DNA in strains of X. campestris 34 

4-3 Hybridization of radiolabeled plasmid probes to 

total DNA of pathovars of X. campestris 44 

5-1 Strains of X. campestris pv. malvacearum from cotton 

used in race-specificity investigation 52 

5-2 Plasmid groupings and cotton plant reactions to strains of 

X. campestris pv. malvacearum 55 

5-3 Pathogenicity of X. campestris pv. malvacearum 

transconjugants on cotton host differentials 59 

5-4 Selection of plasmid replication genes in X. campestris 

pv. malvacear'jm strain X 7. . 7777777777 62 

6-1 Strains of X. campestris used for restriction fragment 

length polymorphism analysis 73 

6-2 Sizes of DNA fragments from X. campestris genomic digests 

which hybridized to the XCTl DNA probe. See Figure 6-3 83 

6-3 Sizes of DNA fragments from X. campestris genomic digests 

which hybridized to the XCTl DNA probe. See Figure 6-5 87 



tri 



LIST OF FIOJRES 



Figure 



Page 



3-1 SDS-Polyacrylamide gel electrophoresis of total proteins 

from strains of X. campestris 23 

4-1 Plasmid DNAs from strains of X. campestris pv. malvacearum 

digested with restriction endonucleases EcoRl and BamHI 37 

4-2 Graphic representation of plasmid EcoRl restriction fragment 

profiles for pathovars of X. campestris 38 

4-3 Plasmid DtiAs from strains of X. campestris pv. citri 

digested with restriction endonuclease EcoRI 39 

4-4 Plasmid DNAs from strains of X. campestris pv. 

malvacearum digested with restriction endonuclease 

EcoRI and hybridized to probe N4 . 5 41 

4-5 Plasmid DNAs from strains of X. campestris digested with 
restriction endonuclease EcoRI and hybridized to orobes 
P2.0 and P2.3 '; 43 

5-1 Plasmid curing of X. campestris pv. malvacearum with 

SDS treatment 50 

5-2 Transconjugants of X. campestris pv. malvacearum containing 

pLXD and hybridized to the plasmid probe pSa4 64 

5-3 Transfonriation of E. coli ED8767 with plasmid DNA 
frcm X. campestris pv. malvacearum transconjugants 
mated with pLXD gg 

5-4 Plasmid DNA of X. campestris pv. malvacearum strain X 
digested with restriction endonucleasea and hybridized 
to the plasmid probe pXD-1 67 

6-1 Genomic DNA of strains of X. campestris pvs. phaseoli , 
alfalfae , and campestris digested with restriction 
endonuclease Eco RI 78 

6-2 Genomic DNA of strains from different pathovars of 

X. campestris digested with restriction endonuclease EcoRI .. 79 



Vll 



6-3 Genomic DNA of strains of X. cair.pestris pvs. phaseoli, 
alfalfae, and campestris digested with restriction 
endonuclease EcoRI and hybridized with probe XCTl 81 

6-4 Genomic DNA of strains of X. campestris pvs. phaseoli , 
alfalfae , and campestris digested with restriction 
endonuclease EcoRl and hybridized with probe XCTll 84 

6-5 Genomic DNA of strains from different pathovars of 

X. campestris digested with restriction endonuclease EcoRI 

and hybridized with probe XCTl 777 85 



Vlll 



Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 

GENETIC ANALYSIS OF INTRASPECIFIC VARIATION IN 
PATHOVARS OF Xanthcmonas campestris 

By 

Gerard Raymond Lazo 

May, 1987 

Chairman: Dean W. Gabriel 

Major Department: Plant Pathology 

Xanthomonas campestris is always found in association with 
plants. Those strains of X. campestris which are known to be 
pathogenic are differentiated into over 125 pathovars (pathogenic 
variants), on the basis of limited pathogenicity tests. Sane members 
of a pathovar may have a broader host range than others, since 
exhaustive pathogenicity testing is impractical. Other than these 
tests, there is no definitive means to classify an unknown X. 
campestris isolate. Those strains of X. campestris which are not 
pathogenic are unclassifiable by the pathovar system, yet they may 
exhibit similar host range specificity to those which are pathogenic. 
This work was conducted in an effort to better understand the 
variability of x. campestris . 

Over 115 strains of X. campestris were examined for plasmid 
content and restriction endonuclease profiles. Of the 26 pathovars 
examined, only 13 V7ere found to contain plasmids. Restriction 



,ix 



endonuclease digested plasmid DNA from strains within a given 
plaanid-containing pathovar gave surprisingly similar, but not always 
identical, digestion profiles by agarose gel electrophoresis. All 60 
strains tested of X. campestris pvs. glycines , malvacearum , phaseoli , 
and vignicola could be accurately identified by pathovar fran 
determination of the restriction fragment profile and/or by Southern 
hybridizations of that profile with a selected plasmid DNA probe. In 
no instance was the same plasmid profile seen in more than one 
pathovar- The apparent stability of the plasmids provides a natural 
genetic marker that can be strain specific and perhaps useful in 
epidemiological investigations. 

Cloned DNA fragments, derived from a cosmid library of a Florida 

isolate of X. campestris pv. citri strain 3041, were used to detect 

restriction fragment length polymorphian (RFLP) of total DNA from 87 

strains of X. campestris , comprising 23 different pathovars. 

Autoradiographs of Southern transfers of genomic DNA probed with 

cosmid-sized clones revealed hybridization profiles which appeared to 

be highly conserved. The hybridization patterns observed between 

different pathovars suggested that some DNA fragments were conserved 

at the species level, others were conserved at the pathovar level, and 

still others were variable. The degree of variability appeared to 

depend on the DNA probe used. By using more than one DNA probe, or by 

digesting the genomic DNAs with different restriction endonucleases, 

all of the strains of X. campestris described as belonging to a given 

pathovar could be differentiated. All strains of X. campestris vjere 

readily grouped by RFLP phenotypes, and the classification based on 



RFLP patterns correlated well with the classification based on 
pathogenicity. Although certain pathovars may need to be redefined, 
tliis work supports and helps validate the natural taxonomic groupings 
provided by the pathovar naming system. 



XI 



CHAPTER WE 
INTRODUCTION 



The genus Xanthononas Dowson consists of bacteria described as 
being gram-negative, obligately aerobic, non-fermentative rods which 
are motile by a single polar flagellum and can produce a broninated 
yellow pigment, called xanthcmonadin (6) . All reported strains of 
this genus have been described as plant associated, and most are 
reported as being pathogenic to a particular plant host. Based on 
microbiological classification, this genus can be separated into at 
least five separate species: X. campestris (Paimiel) Dowson, X. 
albilineans (Ashby) Dowson, X. ampelina Panagopoulos, X. axonopodis 
Starr and Garces, and X. fragariae Kennedy and King (and possibly X. 
PPP^^^ (Ride) Ride and Ride). Over 125 variants, or pathovars, of X. 
campestris , are essentially indistinguishable from each otiier, except 
for their plant host range (6,89). 

To maintain order in differentiating among strains of X. 
campestris by host range, the rank of pathovar has been assigned in 
addition to the species name (24,98). The primary means by which 
strains can be differentiated as pathovars is by inoculation of the 
plants which serve as susceptible hosts to a given pathovar. The 
convention for naming pathovars of X. campestris has generally been to 
name it for the host from which it was first isolated (89) . A 
particular pathovar designation suggests a limited host range for a 

1 



given strain. The actual host range may be much broader, but a 
complete host range description of a given strain would require 
testing on all plants known to support growth of the genus. This 
would be a monumental task, and therefore the potential host range of 
a given strain of X. campestris is unknown. Obviously, any 
information regarding pathogenicity is of value to pathologists. 
However, without knowledge of the potential host range variation of 
strains within a pathovar, the taxonaric value of the pathovar system 
is questionable. 

Attempts to devise alternative means to differentiate pathovars 
of X. campestris have included serology (1,94), membrane protein 
profiles (70) , phage-typing (37) , rRNA-DNA and DNA-DNA hybridization 
(21,73), and gas chromatography of fatty acids (65,81). Although most 
of these methods have been useful to differentiate given sets of 
strains of X. campestris , there is little evidence that the groupings 
distinguished by these methods have any correlation to the host range 
groupings observed for given pathovars of X. campestris . None of 
these techniques have been successful in replacing pathogenicity tests 
to identify X. campestris to the pathovar level. Because differences 
in host range primarily determine the pathogenic variability observed 
in X. campestris , it is of interest to know more about the nature of 
this microorganism and the mechanisms which determine host-ranqe 
specificity. 

Genetic analyses of pathogenicity can easily be done with 
bacteria, for they can be quickly cloned and assayed, and studied 
using current recombinant DNA technologies (50) . X. campestris is 



useful because it is a facultative parasite with a high level of 
natural variation and has been shown to follow similar gene-for-gene 
patterns of interactions with plants as found with fungal pathogens 
(33). 

The objective of this research was to develop a better 
understanding of the intraspecific variation and specificity present 
in the species X. campestris . This was conducted by first verifying 
the pathogenicity of the working strains as pathovars and, where 
appropriate, as races within selected pathovars of X. campestris 
through pathogenicity testing. This work was followed by physical 
characterization of the plasmids and chromosomes of x. campestris 
races and pathovars. These DNA analyses were designed to determine 
the extent of physical variation within the species, pathovars, and 
races of X. campestris . Finally, genetic variation was analyzed at 
the race level by using cloned genes and complementation analyses to 
identify specific avirulence genes which contribute to the race 
phenotypes. By conducting experiments in a multi-faceted approach, it 
was hoped that a better understanding of the interactions of strains 
of x. campestris with their host plant (s) could be appreciated. 



CHAPTER TWO 
LITERATURE REVIEW 



At least 124 monocotyledonous and 268 dicotyledonous plants 
species are susceptible to infection by bacterial strains of the genus 
Xanthomonas (57). Compiled listings indicate that the genus 
Xanthononas has a wide host range extending over nine monocotyledonous 
families comprising 66 genera, and 49 dicotyledonous families 
comprising 160 genera. Pathogenicity toward gymnosperms or lower 
plants has not been reported. 

Although pathovars are generally assumed to have a restricted 
host range, seme strains of X. campestris have extended host ranges to 
more than one plant species. in these instances, host range usually 
extends toward other members of the same plant family (with few 
exceptions) . For example, the plant families Compositae, Cruciferae, 
Graminae, and Leguminosae each contain different species susceptible 
to the same X. campestris pathovar (X. c. pv.) . An example of a broad 
host range pathovar of graminaceous hosts is X. c. pv. translucens 
(Jon., John., and Redd.) Dye, which is pathogenic on barley, rye, and 
wheat. Other strains pathogenic on graminaceous hosts have been 
described, but their interrelationships among the other pathovars were 
poorly defined (6), In another example, X. c. pv. alfalfae , is 
pathogenic on the legumes alfalfa, bean, and pea. Another special 
relationship is that of X. c. pv. vesicatoria (Doidge) Dye, which is a 

4 



pathogen of Capsicum annuum L. (pepper) and Lycoperslcon esculentum 
(tomato). While some strains of X. campestris appear to have a host 
range to more than one plant species, most others appear to have an 
extremely narrow host range (6,57). The host range characteristic is 
considered to be a stable trait, although there is one report of 
altered host range by artificial selection (23) . This report has been 
refuted, with no change in host-range specificity observed for X. 
canpestris (84). Within at least some pathovars, variability among 
strains exists as race-specific plant interactions (8,17,20). These 
are recognized when resistant plants of an otherwise susceptible plant 
species are discovered. The ability to differentiate races of x. 
campestris pathovars is dependent on the discovery of plant hosts, 
called host differentials, which exhibit variability in their response 
to different pathogenic strains. Pathogenic races have been described 
for X. c. pv. malvacearum (Snith) Dye, X. c. pv. oryzae , X. c. pv. 
translucens , X. c. pv. vesicatoria, and possibly others. This race- 
specific variation in X. campestris was suggestive of gene-for-gene 
interactions between pathogenicity determinants in the pathogen with 
resistance genes in the host (8,20). 

Gene-for-gene relationships can only be established by genetic 
crosses of the host and the parasite. Because the bacterial genetics 
can only be done with recombinant DNA techniques, gene-for-gene 
interactions could only be recently established. Proof of 
gene-for-gene interactions with bacterial pathogens was documented for 
X. c. pv. malvacearum (33) . 



The high level of host-range and race specificity described above 
for X. campestris is observed with some other plant pathogenic 
bacteria. For example, there are more than forty pathovars of 
Pseudcmonas syringae which are primarily distinguished by host-range 
specificity (75,86). Seme P. syringae pathovars have host ranges 
which are narrowly restricted to specific plant species; others have a 
wider range of hosts. Strains of Erwinia chrysanthemi are currently 
being considered for differentiation at the pathovar level (56) . 
Strains of Agrobacter ium spp. and Rhizobium spp. are not 
differentiated using the pathovar system; however, host-specific 
interactions analogous to those observed in X. campestris exist with 
these bacteria. Species within each genus can be differentiated 
biochemically and by plant inoculations. These two genera, which 
cause very different symptoms, are both in the family Rhizobiaceae and 
are genetically highly related (25,42). In Rhizobium spp. and 
Agrobacterium spp., host range appears to be specified by the 
bacterium by genes which are expressed as a positive function (32) . In 
several species of Rhizobium , host range function has been identified 
as being plasmid borne (4). The introduction of one of these plasmids 
may extend the host range of a limited host range species of 
Rhizobium . 

Similarly, A. tumefaciens tumor inducing (Ti) plasmids carry host 
range determining genes (64) . Ti plasmids from wide-host-range 
strains, when moved to A. tumefaciens strains with limited host- 
ranges, confer a wide host range to the recipient (10,43,48). The 
pim plasmid, found in A. tumefaciens strains from grapevine, carries 



a gene for tartaric acid catabolism (35) . The expression of pTAR 
genes appears to function more like a host range determinant rather 
than an oncogenic determinant, as found on tumor inducing plasmids. 

The proof of plasmid involvement in virulence and host range with 
species of Agrobacterium and Rhizobium apparently stimulated research 
of plasmid involvement among other phytopathogenic bacteria. Plasmids 
have been found to exist in many other phytopathogenic bacteria, but 
few have been associated with pathogenicity (76) . Examples of plasmid 
functions in pathogenicity on plants include tumor induction in A. 
tumefaciens (97) , nodulation in Rhizobium spp.' (62) , indoleacetic acid 
production by P. syringae pv. savastanoi (90) , cytokinin production by 
Corynebacterium fascians (syn. Rhodococcus fascians ) (72) , virulence 
of Pseudomonas solanacearum (5) and race-specific avirulence in X. c. 
pv. vesicatoria (88), among others. A general conclusion from these 
studies is that plasmids from a v/ide range of plant pathogenic genera 
can carry a variety of genes which determine the outcome of plant- 
pathogen interactions. 

Plasmids have been shown to be physically similar among strains 
within pathovars of p. syringae (77) . The similarity is base3 on 
hanology studies using restriction endonuclease digestion patterns and 
DNA-DNA hybridization experiments (15) . Additionally, there are 
instances where similar plasmids, or even identical ones, may be found 
in closely related pathovars of P. syringae (77). Although the 
functional utility of these plasmids is unknown, they often cannot be 
eliminated or "cured" from the bacterial strain, possibly because such 
an event would be lethal. For example, in P. syringae pv. 



phaseolicola , a plasmid, pMCTlOS, can be forced to integrate into the 
host chromoscme, but cannot be cured (79,92). The plasmid pfCVlOB has 
sane hanology with other P. syringae pv. phaseolicola strains, and 
even P. syringae pv. glycinea strains. Homology between these 
plasmids may be due to transpositional events, the presence of 
repetitive sequences, variability due to integration and excision of 
the plasmid, plasmid origin of replication, or other plasmid 
maintenance function (79,92). Another possibility is that genes 
essential for pathogenicity, host-range, or race specificity may be 
encoded on them. The association of plasmid homology with host-range 
specificity may indicate the involvement of plasmids in host range 
determination in these cases. 



CHAPTER THREE 
HOST RANGE OF PATHOVARS OF XANTHOMONAS CAMPESTRIS 



Introduction 
There are over 125 pathovars of X. campestris described, and the 
only means to differentiate between them is by pathogenicity testing 
(6) . Seme strains of X. campestris have extended host ranges to more 
than one host, usually to other members of the same plant family, with 
few exceptions (57) . The host range can sonetimes overlap onto 
another host for which a different X. campestris pathovar has been 
described. For instance, X. c. pv. alfalfae is a pathogen of the 
legumes alfalfa, bean, and pea, but strains exist of x. c. pv. 
phaseol i and X. c. pv. pisi which have host ranges limited to bean and 
pea, respectively. The relatedness between these strains at the 
subspecific level is unknown, except for their host ranges. In 
another instance, strains of X. c_. pv. malvacearum have been isolated 
from the malvaceous hosts, cotton and hibiscus (7,12). Although given 
the same pathovar designation because they were each isolated fran 
plants of the Malvaceae family, their relatedness is unknown. Similar 
situations, where the relationship between strains is unclear, exist 
among other pathovars of X. carripestris (6) . The purpose of this 
investigation was to verify the pathogenicity and host range of all 
stock strains used in this dissertation, to examine the relationships 
between apparently related strains of X. campestris in terms of host- 

9 



10 



range specificity, and to find a physiological test diagnostic for one 
or more pathovars used in tliese studies. 

Materials and Methods 
Bacterial Strains and Host Plants 

Bacteria used in this investigation consisted of pathovars of X. 
campestris which were pathogenic to members of either the Leguminosae 
or Malvaceae plant families. The pathogens of leguminous host plants 
consisted of X. campestris pvs. alfalfae, cvamopsidis, glycines, 
phaseoli , pisi, and vignicola isolated from alfalfa, guar, soybean, 
kidney bean, pea, and cowpea, respectively. The pathogens of 
malvaceous host plants consisted of strains of X. c. pv. malvacearum 
isolated from cotton or hibiscus. The strains of X. campestris used 
in this study are shown in Table 3-1. The legume host plants used in 
this investigation included Medicago sativa cv. "FL-77", Glycines max 
cv. "Evans", phaseolus vulgaris cv. "California Light Red", and yigna 
unguiculata cv. "California Blackeye #5". The malvaceous hosts 
included three cultivars of Hibiscus rosa-sinensis and eight cultivars 
of Gossypium hirsutum . One of the cultivars of H. rosa-sinensis was 
cv. Brilliant Red; the other two were not named, but chosen to 
represent diversity in the species. Eight cotton cultivars were 
selected for differentiating races of X. c. pv. malvacearum , two of 
which were cv. 101-102B, and cv. Gregg (45) . The other cultivars were 
cv. Acala 44 (no resistance to cotton strains) , and five cultivars 
derived from cv. Acala 44 crossed to obtain single gene resistance to 
the cotton strains of X. c. pv. malvacearum , the cultivars being Acala 
B^, B2, B3, B5, and Bjj^ (7) , 



11 



Table 3-1. Strains of X. campestris used in host range investigati 



gation. 



Pathovar Strain 



Plant Host 



Location 



glycines 



malvacearum 



alfalfae KS 
FL 
cyamopsidis 13 D5 

X002,X005,X016,X017 

B-9-3 

1717 

17915 

S-9-8 

D,M,N,0,U,V,X,Y,Z,TX84 

A,B,E,F,G,H 

Chl,Ch2 

HV25 

Su2,Su3 

FL79 

083-4244 ,M8 4-11 

X10,X27,X52,X102,X108 

EKll,Xph25,Xpfll 

Xpa,Xpll 

82-1,82-2 

LB-2,SC-3B 

XP2 

XP-JL 

XP-JF 

XP-DPI,B5B 

XPl 

A81-331,C-1,CB5-1, 

Xvl9,SN2, 432, 82-38 



phaseoli 



pisi 
vignicola 



Medicago sativa 

M. sativa 

Cyamopsis tetragonoloba 

C. tetragonoloba 

Glycine max 

G_. max 

G . max 

G. max 

Gossypium hirsutum 

G. hirsutum 

G . hirsutum 

G. hirsutum 

G - hirsutum 

G. hirsutum 

Hibiscus rosa-sinensis 

H_. rosa - sinensis 

Phaseolus vulgaris 

P_. vulgaris 

P . vulgaris 

P_. vulgaris 

P . vulgaris 

P. vulgaris 

P. vulgaris 

P. 



vulgaris 
Pisum sativum 
Vigna ungiuculata 
V. unguiculata 



Kansas 
Florida 

Arizona 

Brazil 

Africa 

Wisconsin 

Texas 

Oklahoma 

Chad 

Upper Volt a 

Sudan 

Florida 

Florida 

Florida 

Nebraska 

Wisconsin 

Florida 

Nebraska 

New York 

Kansas 

Missouri 

Japan 

Georgia 

Georgia 



12 



Plant Inoculations 

Pathogenicity tests were conducted by pressure infiltrating 
bacterial suspensions into leaf tissue with a blunt-end syringe and 
incubating the plants in 28 C or 30 C growth chambers until symptons 
were expressed. The leguminous host plants were incubated at 28 C and 
symptcms were recorded after one week. The malvaceous hosts were 
incubated at 30 C and symptoms were recorded after a two week period. 
Pressure infiltration facilitated rapid screening of plant reactions, 
but was not representative of natural inoculation. Because alfalfa 
leaves were small in size, infection was by spray inoculation. 
Bacterial suspensions were prepared by centrifuging overnight cultures 
of X. campestris , resuspending them in 0.7% NaCl to an approximate 
optical density (ODgQOnm) of 0-3 for plant inoculation. These 
pathogenicity tests were repeated at least once. The results were 
recorded to determine compatible (pathogenic) or incompatible (non- 
pathogenic or hypersensitive) plant reactions to inoculations. 
Physiological Differentiation 

As X. campestris is defined (6), the physiological 
characteristics for gelatin and starch hydrolysis may be variable for 
this species. Other physiological characteristics which are consistent 
with this species includes xanthomonadin production, mucoid growth, 
and esculin and casein hydrolysis. These above mentioned 
characteristics were determined using standard methods (36,82). 
Additionally, physiological tests for production of cellulase (96), 
lecithinase, lipase, and pectinases were included. Hildebrand's 
medium prepared at three pH's (4.5, 7.0, and 8.5) was used to detect 
pectolytic enzyme activity. in each of these three media, sodium 



13 



polypectate was used as the carbon source. The pH 5 medium included 

only pectin as the sole carbon source. These tests were each repeated 

once. 

Polyacrylamide Gel Electrophoresis 

Bacteria were grown overnight at 30 C in a peptone-glycerol 
broth. Cells were adjusted to a 0.3 optical density (ODgQOnm) ' ^^'^ 
1.5 mis of cells were collected by centrifugation, washed once in 
water, and resuspended in 50 ul 10% sorbitol. Then, 50 ul 
solubilization buffer (90 mM Tris (pH 6.8), 20% glycerol, 4% SDS, 10% 
B-mercaptoethanol, and .00 2% bromophenol blue) was added to the 
suspension and the mixture was boiled for 5 minutes. Approximately 10 
mg protein (10 ul solution) was added to a 10% resolving acrylamide 
gel (51), and samples were electrophoresed in a 20 cm x 17 cm x 1.5 mm 
vertical gel unit at 10 V/cm and 15 C until the dye reached the bottom 
of the gel. Biorad low-range SDS-PAGE standards were used as 
molecular weight markers. Gels were stained with 0.1% Coonassie blue 
in 42% methanol and 17% glacial acetic acid. The gel was destained in 
30% methanol and 10% glacial acetic acid. Electrophoresis was only 
performed once. A photograph of the gel was taken using Polaroid type 
55 film. 

Results 
Pathogenicity Tests 

All of the strains of X. c. pv. phaseoli , with one exception, 
gave the expected disease reactions on the legume hosts tested. The 
exception, strain B5B, was found to be non-pathogenic to all four 
hosts tested. The remaining six strains were pathogenic on kidney 
bean (Table 3-2) , giving watersoaked lesions on inoculated leaves 



14 



Table 3-2. Legume plant reactions to inoculation with pathovars of X 
campestris. — " 



Inoculation Strain 



X. campestris pv. 



bean 



Host Reaction^ 



cowpe^ 



soybean 



alfalfa 



phaseoli 


82-1 


+B 




Xpa 


+ 




Xpfll 


+ 




XP-JF 


+ 




EKll 


+ 




XP2 


+ 




B5bG 


- 


vicrnicola 


CB5-1 


_H 




Xvl9 


.. 




432 


— 




A81-331 


_ 




C-1 


_ 




82-38 


-f/-K 




SN2 


+/- 


glycines 


&-9-3 


+L 




1717 


+ 




17915 


+ 




S-9-8G 


- 


alfalfae 


EL 


+L 


malvacearum 


N 


_F 


control 








_c 



D 



A = + is compatible, - is incom- 
patible, +/- is intermediate, 
and is a null reaction. 

B = compatible lesions were 
watersoaked and appeared to 
be spreading. 

C = dry necrotic lesion with v/ine 
red reaction. 

D = dry necrotic lesion with 
chlorosis. 

E = no reaction seen with spray 
inoculation. 

F = slight tissue discoloration 
at inoculation site. 



_F 



+1 

+ 
+ 
+ 
+ 
+ 
+ 








_J 



+ 
+ 



.D 
F 

























N 



G = strain appeared non- 
pathogenic. 

H - dry necrotic lesion with 
slight watersoaking at 
periphery of inoculation site. 

I = dry necrotic lesion with 
shothole effect. 

J = dry necrotic lesion. 

K = as described for H, but 

slight shothole effect present. 

L = watersoaked lesion. 

M = watersoaked chlorotic lesion. 

N = watersoaked leaf soots. 



15 



after four days. The lesions extended from the inoculation sites 
giving the appearance that the bacterium was spreading throughout the 
leaf. 

The strains of X. c. pv. phaseoli which were found pathogenic on 
kidney bean were not found to be pathogenic on cowpea, soybean, or 
alfalfa. In cowpea and soybean, a hypersensitive reaction was 
apparently elicited, resulting in dry necrotic lesions at the 
inoculation sites. In soybean the lesion was slightly chlorotic, 
whereas for cowpea, a wine red lesion occurred. However, X. c. pv. 
phaseol i strain XP2 did not elicit the wine red color on cowpea (Table 
3-2) . No response was observed on alfalfa from the spray inoculations 
with X. c. pv. phaseoli . 

The strains of X. c. pv. vignicola were pathogenic on cowpea, 
resulting in irregularly shaped lesions which had a tearing, or 
"shothole" appearance at the inoculation site. On soybean, the 
inconpatible response appeared dry and collapsed. Again, no response 
was observed from spray inoculation on alfalfa (Table 3-2) . 

On kidney bean, a range of different responses occurred from 
inoculations with X. c. pv. vignicola . For two of the strains, A81- 
331 and CI, an incompatible response appeared as a dry collapsed 
lesion. For the other strains of X. c. pv. vignicola , there appeared 
to be seme slight water-soaking in the tissue about the periphery of 
the inoculation site. In two of tliese instances, for strains 8238 and 
SN2, the "shothole" effect as seen on cowpea was observed. 

The strains of X. c. pv. glycines appeared to be pathogenic on 
both soybean and kidney bean, but not on cowpea or alfalfa (Table 3- 
2). The pathogenic responses appeared as watersoaked lesions, with 



the lesion being slightly chlorotic in soybean. The incompatible 
response in cowpea appeared as wine red in color like that seen with 
X. c. pv. phaseoli . 

From spray inoculations on alfalfa, only x. c. pv. alfalfa , 
resulted in a pathogenic response. After 7 days, small water-soaked 
leaf spots which later turned necrotic appeared. X. c. pv. hlfalfa 
was also pathogenic on kidney bean, and appeared so slightly on 
soybean. There were small water-soaked spots which occurred at the 
inoculation sites on soybean. The wine red incompatible response was 
observed on cowpea. 

A strain of X. c. pv. malvacearum , which was not a pathogen of 
legumes, resulted in only a slight tissue discoloration at the 
inoculation site on bean and soybean. A null reaction, as observed 
with the water-only control, was observed on cowpea. The water-only 
control also resulted in a null reaction on the other hosts tested. A 
summary of the plant reactions observed for the inoculation 
experiments is given in Table 3-3. 

In separate inoculation tests, a strain of X. c. pv. cyamopsidis '" 
was not found pathogenic on alfalfa, kidney bean, cowpea, or soybean 
and a X. c. pv. pisi strain vras not found pathogenic on kidney bean. 
Overall, most of the reactions observed conformed to those reported in 
the literature, with some exceptions. Variation was evident among 
strains of a given pathovar in addition to that between pathovars. 
The X. c, pv. alfalfae strains appeared to have overlapping host 
ranges, which extended to bean and pea in addition to alfalfa. The 
kidney bean cultivar, California Light Red, appeared susceptible to a 



16 



Table 3-3. Host range observed for pathovars of Xanttiomonas campestris 
pathogenic to legume host plants. 

Inoculation reaction on host plants^ 

Pathovar G.max M.sativa P. sativum P. vulgaris V.unquiculata 



17 



alfalfaeb 


,^ 


glycines 


+ 


phaseoli 


_e 


pisi 


_. 


vignicola 


_e 


cyamopsidis^ 


- 



+^ + jS. 

nt + _ 

nt^ + _ 

+^ - 

nt +g + 

nt ~ 



+ - compatible, - = incompatible, nt = not tested. Host plants were 
Glycines max (soybean) cv. Evans; Medicago sativa (alfalfa) cv. 
FL-77; Pisum sativum (pea) cultivar not known; Phaseolus vulgaris 

:, (bean) cv. California Light Red; and Vigna unguicu lata (cowpea)' cv. 

,;! California Blackeye #5. 

1 , 

Other susceptible hosts reported are Trigonella and Mel i lotus (6) . 

^ Reported as susceptible, but not tested. 

, ° Reported as positive in the literature (6) . 

^ Reaction appears negative for most strains, but occasional limited 
water soaking is evident. 

1 There is a report of pathovar phaseoli on Pisum lunatus (6) . 

5 Symptons are similar to those seen in V. unguiculata . 
^ The natural host belongs to the genus Cyamopsis . 



18 



representative strain from each of the X. canpestris pathovars of 
legumes tested, with the exception of X. c. pv. cyamopsidis . Various 
phenotypic responses were observed on the cotton and hibiscus 
cultivars with the X. c. pv. malvacearum strains tested. In general, 
it appeared that those strains of X. c. pv. malvacearum derived as 
pathogens from cotton were pathogenic to one of the hibiscus cultivars 
tested. None of the X. c. pv. malvacearum strains derived as 
pathogens from hibiscus were pathogenic on the cotton cultivars in 
these tests. 

From the inoculation tests on the cotton host differentials with 
the cotton derived strains of X. c. pv. malvacearum , it appeared that 
three races of the pathogen were being used. All of these strains 
were pathogenic on Acala 44 (susceptible host) , but could be 
differentiated into races by the other cotton lines containing 
different resistance gene backgrounds. The X. c. pv. malvacearum 
strain N was pathogenic on all the cotton lines tested, while strain H 
was only pathogenic on Acala 44. Strains FL79 and TX84 appeared to be 
the same race because they were pathogenic on the same six out of 
eight cotton lines tested (Table 3-4). However, these two strains 
differed in reactions on the three hibiscus cultivars inoculated. 
Strain FL79 appeared pathogenic on the three cultivars, but TX84 gave 
a null response similar to the water-only control. The two other X. 
c. pv. malvacearum strains (N and H) api^ared identical in reaction 
with the three hibiscus cultivars,. giving pathogenic responses to 2 
cultivars, and a null response on the remaining cultivar (Table 3-4) . 



19 



Table 


3-4. 


Malvaceous 


host plant reactions 


to 


inoculation 


with 


strains 


of X. 


::ampestris 


pv. malvacea 


rum^. 














Strain 










Cotton 










Hibiscus 








Acala 






101 


Gregg 


Hl^^ 


H2 




44 


% 


B2 


B3 


B5 


^IN 


H3 


N 


+ 


+ 


+ 


+ 


+ 


+ 


+ 


+ 


+ 





+ 


H 


+ 


- 


- 


- 


- 


— 


_ 


_ 


+ 





+ 


n.79 


+ 


+ 


+ 


+ 


+ 


_ 


_ 


+ 


+ 


+ 


+ 


TX84 


+ 


+ 


+ 


+ 


+ 


_ 


_ 


+ 











XIO 


- 


- 


- 


- 


- 


_ 


_ 


_ 


+/- 


+ 


+ 


:' X27 



































X52 


_ 


_ 
































— 


^ 


— 


— 


+ 


+ 


XI 02 


_ 


_ 


































— 


— 


— 


+ 


+ 


X103 


_ 


,^ 








































+ 


+ 


X108 


- 


- 


- 


- 


- 


_ 





^ 


+ 


+ 


+ 


I 83-4244 


- 


- 


- 


- 


- 


- 


_ 


_ 




+ 


+ 


MB 4-11 


























n 








84-1093 

( 

I 


^0 





—" 



















+/- 






° + - compatible reaction, - = incompatible reaction, = null 
response, and +/- = intermediate reaction. 

" K. rosa-sinensis cv. "Brilliant Red". 

^ Strain 84-1093 is X. c. pv. esculenti. 



20 



With the eight X. c. pv. malvacearum strains derived from 
hibiscus, two gave null reactions on all of the cotton and hibiscus 
cultivars tested (Table 3-4) . The other strains derived from hibiscus 
were pathogenic on at least two of the three hibiscus cultivars 
tested, but were not pathogenic on any of the cotton cultivars. in a 
separate experiment, one of the strains (M84-11) did appear to give 
pathogenic symptoms on Acala 44, while another strain (83-4244) gave a 
hypersensitive response, in this instance the mean temperature was at 
25 C, rather than 30 C. 

An additional strain tested, X. c. pv. esculenti strain 84-1093, 
which was a pathogen derived from okra, gave a null reaction on seven 
of eight cotton cultivars and one of three hibiscus cultivars. The 
other cotton or hibiscus cultivars gave hypersensitive responses 
(Table 3-4) . 
Physiological Differentiation 

All of the strains tested were mucoid on NYGA medium, produced 
xanthcmonadin, and hydrolyzed esculin and casein. All of the strains 
tested also appeared to be positive for starch and gelatin hydrolysis, 
and production of lipase (Table 3-5) . Six of the seven X. c. pv. 
vignicola strains, and one strains of X. c. pv. phaseoli appeared to 
hydrolyze gelatin slower than the other X. campestris strains. 
Cellulase activity, indicated by pitting of the solid medium, was 
observed for all seven of the pathovars tested. However, not all 
strains of X. c. pv, phaseoli appeared to have cellulase activity, and 
X. c. pv. alfalfae strains were weaker in this activity. None of the 
strains tested exhibited pectolytic activity at pH 4.5, suggesting a 



21 



Table 3-5. Physiological reactions of strains of X. cair.pestris, 





















Pathovar 


Starch 


Physioloaical 
N 
Gelatin Cellulose 4 


Test Med. 
a polypec 
.5 7.0 


Lum^ 






:tate 
8.5 


Pectin 
5.0 


Lecithin^ 


alfalfae 


















KS 


+ 


+ 


w 


_ 


_ 






+ 


FL 


+ 


+ 


w 


_ 


^ 






+ 


cyamopsidis 


















13D5 
glycines 


+ 


+ 


+ 


- 


+ 


+ 


- 


+ 


B-9-3 


+ 


+ 


+ 


^ 


+ 


+ 




+ 


1717 


+ 


+ 


+ 


_ 


+ 


+ 




+ 


17915 


+ 


+ 


+ 


_ 


+ 


+ 




+ 


S-9-8 
malvacearum 
H 


+ 


+ 


+ 


- 


+ 


+ 


- 


+ 


+ 


+ 


+ 


_ 










N 


+ 


+ 


+ 


_ 










phased i 


















EKll 


+ 


+ 


— 


^ 


_ 






+ 


Xph25 


+ 


+ 


+ 


^ 








+ 


Xpfll 


+ 


+ 


+ 


_ 








Xpa 


+ 


w 


— 


__ 


— 






+ 


Xpll 
82-1 


+ 
+ 


+ 
+ ■ 


— 


- 


- 


- 


- 


+ 
+ 


82-2 


+ 


+ 


— 


— 


_ 


_ 




+ 


LB- 2 


+ 


+ 






















~~ 


^ 


^ 


^ 


+ 


SC-3B 


+ 


+ 


+ 


_ 


+ 


+ 






XP2 


+ 


+ 






















~~ 


__ 


•— 


^ 


+ 


XP-JL 


+ 


+ 


- 


— 


_ 


__ 




+ 


XP-JF 


■f 


+ 


— 


_ 


-_ 






j- 


XP-DPI 


+ 


+ 






















"" 


— 


— 


^ 


+ 


pisi 


















XPl 


+ 


+ 


+ 


_ 


+ 


+ 




+ 


vignicola 


















A81-331 


+ 


+ 


+ 


_ 


+ 


+ 




+ 


1 C-1 


+ 


w 


+ 


— 


+ 


+ 




+ 


' CB5-1 


+ 


w 


+ 


_ 


+ 


+ 




+ 


Xvl9 


+ 


w 


+ 


_ 


+ 


+ 




+ 


SN2 


+ 


w 


+ 


— 


+ 


+ 




+ 


432 


+ 


w 


+ 


_ 


+ 


4- 




+ 


82-38 


+ 


w 


+ 


— 


+ 


+ 


- 


+ 



^ All strains were mucoid, esculin positive, milk proteolytic, and 
lipase positive. + = positive, - = negative, and w = weak reaction. 

A lecithinase positive reaction was observed as a clearing zone around 
tne colony plus precipitation in the medium. 



22 

lack of polygalacturonase activity. At pH 7.0 and 8.5, pectolytic 

activity was observed for all strains of X. c. pv. cyamopsidis , X. c. 

pv. glycines, X. c. pv. £isi, and X. c. pv. vignicola . Only one of the 

13 X- c. pv. phaseoli strains tested (SC-3B) had pectolytic activity. 

None of the strains degraded pectin as a carbon source, only sodium 

polypectate. 

Polyacrylamide Gel Electrophoresis 

Total protein was extracted from 18 strains, representing seven 
different pathovars of x. campestris . The pathovars included were X. c. 
pv. alfalfae, x. c. pv. cyamopsidis , x. c. pv. glycines , x. c. pv. 
malvacearum, X. c. pv, phaseoli , X. c. pv. pisi, and X. c. pv. 
vignicola. All of the pathovars, except X. c. pv. malvacearum (Fig. 3- 
1, lane P) , are known pathogens of legumes. One of the X. c. pv. 
Phaseoli strains consisted of a bicxrhemical variant of this pathovar, X. 
c. pv. phaseoli var. fuscans (Fig. 3-1, lane J). One of the X. c. pv. 
malvacearum strains was derived from hibiscus, rather than from cotton 
(not shovm) . 

In general, there were few observable differences in the protein 
banding patterns for the X. campestris strains. Seme minor differences 
in banding patterns between different pathovars were observed, but 
variation to tine same extent was also present within a given pathovar. 

Discussion 

The plant reactions observed from the X. campestris host range 
study conformed to expectations, in most cases. However, it was evident 
that pathogenic variation in strains associated with a given pathovar 
could occur, as shown by the race- specific interactions with the X. c. 



23 




Figure 3-1. Polyacrylamide gel electrophoresis of total proteins from 
strains of X. cair.pestris . Lanes A and B, X. c. pv. alfalfae ; lanes C- 
F, X. c. pv. glycines ; lanes C^I, x. c. pv. phaseo li; lane J, X c 
pv. phaseoli var. fuscans; lane K, X. c. pv. pisi, lane L, X. c~* pv ' 
cyamopsiais; lanes f^^O, X. c. pv. vignicola; "^Hd lane P, X c Sv* 
malvacearum (cotton). Protein size is labeled in kilodaltons" ~* 



24 



pv. malvacearum strains. In the legume study, it was interesting to 
note that different hosts gave different phenotypic responses to 
different incompatible pathovars of X. campestris . These different 
incompatible phenotj-pes suggest that different host resistance genes 
may be involved in each of these incompatible interactions. Although 
the hypersensitive reaction is conmonly considered to be the general 
resistance mechanism of plants against bacteria, these data suggest a 
more specific response, unique to each incompatible reaction (33,49). 
NO formal studies on the hypersensitive reaction were implemented, so 
the significance of this observation will require further 
investigation. Differences in the phenotypes which may account for 
this are time of recognition, cell number, electrolyte leakage, or 
other bacteria- or plant-conditioned responses. Likewise, a 
hypersensitive reaction may have cx:curred for instances where a null 
response was observed (27) . 

The results from inoculation tests on the malvaceous hosts 
suggested that those strains of x. c. pv. malvacearum derived from 
cotton are potential pathogens of hibiscus, whereas those derived from 
hibiscus are generally restricted to hibiscus. However, in one 
experiment, it was found that one of the hibiscus strains caused 
watersoaked lesions on Acala 44 when the night temperatures were 
lowered. This observation indicates that same of the interactions may 
be temperature sensitive. Seme resistance genes in cotton are known 
to be temperature sensitive (9) . 

Race-specific incompatibility may account for the varied X. c. 
pv. malvacearum (cotton strains) reactions on hibiscus. it was 



interesting to note that strains FL79 and TX84, both race 16 of X. c. 
pv. malvacearum cotton strains, each differed in reactions on the 
hibiscus differential lines, since Acala 44 elicited a hypersensitive 
response from 6 of 8 hibiscus strains, possibly Acala 44 contains an 
undiscovered resistance gene to the hibiscus derived strains. Those 
strains for which no hypersensitive reaction was induced may have 
escaped host recognition, but lack the necessary pathogenicity 
determinants to incite disease. Avirulence genes do not give 
phenotypic expression in non-parasites (88). Bacterial growth 
kinetics would help determine if the bacteria are multiplying on these 
hosts. This would be useful for the case of x. c. pv. esculenti , for 
which 1 of 8 cotton hosts were incompatible, and 2 of 3 hibiscus hosts 
were incompatible. 

The data from the protein and physiological tests danonstrated 
differences between strains, but were not useful for typing strains by 
pathovar. There was significant variation among strains within a 
given pathovar. Although the use of SDS-PAGE to study total protein 
profiles did not differentiate the pathovars, some work has suggested 
limiting the analysis to just outer membrane proteins to improve 
resolution of strain differences (70) . 

These experiments confirmed pathogenicity of the stock strains 
and clarified inconsistencies in the literature concerning host range. 
The host range studies indicated that conclusions reached may often be 
strain-dependent, and that examination of a large number of strains 
will reveal a different overall picture than an examination of a few 
strains. For example, the plant source of X. c. pv. malvacearum 



25 



26 

determines whether a strain will have the host range listed for the 

pathovar (hibiscus and cotton) or a more limited host range 
(hibiscus). However, the plant source of X. c. pv. alfalfae, does not 
determine whether a strain will have the host range listed. A strain 
which attacks both alfalfa and bean is X. c. pv. alfalfae . A strain 
with a host range limited to bean is X. c. pv. phaseoli. Finally, 
several biochemical methods were evaluated for their potential to 
classify a strain at the pathovar level. Although standard 
microbiological tests and protein gels were not adequate for such 
classifications, an examination of plasmid DNAs (Chapter Four) and 
chromosomal DNAs (Chapter Six) yielded more rewarding results. 



CHAPTER FOUR 

CONSERVATION OF PLASMID DNA SEQUENCES AND 

PATHOVAR IDENTIFICATION OF STRAINS OF XANTOOMONAS CAMPESTRI^ 



Introduction 
More than 125 different pathovars of Xanthcmonas campestris 
(Pamnel 1895) Dowson 1939 are currently recognized (6,24), and the 
primary means for differentiating them is by inoculation of the plant 
host(s) of that pathovar. It would be a difficult task to inoculate 
every plant that could serve as a host to an X. cam^^stris isolate, 
and therefore the potential host range of a given isolate is unknown. 
Most often, the pathovar name assigned to a strain of X. campestris is 
determined by the host it was isolated fron. Such designations may be 
arti factual since the primary host may be different from the one the 
strain was isolated from; sane X. camj^stris pathovars are known to be 
pathogenic on more than one host. Epiphytes cannot be classified in 
this pathovar identification system. Possible taxonomic relationships 
among pathovars are also elusive. It would be helpful if alternative 
means to differentiate among X. camg^stris pathovars were available. 
Sane suggested approaches for differentiating x. campestris pathovars 
have included serology (1,94), membrane protein profiles (70), 
phage-typing (37), and gas chromatography of fatty acids (65,81). 
These approaches suffer because they are often strain specific. 



Itl i, f contains copyrighted material from the journal 
Phytopathology. It is reprinted here with permission of the publisher 

27 



28 



dependent on constant environmental parameters, and/or so cumbersome 
that no extensive evaluative tests have been performed. 

Plasmid DNA has been identified in several pathovars of x. 
campestris (13,16,30,39,46,52,53,60,80). It is relatively simple to 
extract large numbers of strains and visualize their plasmids with 
standardized alkaline lysis procedures (46,66). To characterize the 
plasmids, restriction endonucleases are used to digest the DNA into 
distinct fragments that can be separated by size, resulting in 
fragment patterns visualized by agarose gel electrophoresis. To date 
there have been no systematic attempts to examine the extent of 
plasmid variation among a large number of strains involving a large 
number of X. campestris pathovars. Our preliminary studies on 
selected pathovars of X. campestris suggested that there was a 
surprisingly high degree of plaanid sequence conservation within some 
pathovars (31,52,53). These studies further suggested that plasmids 
of purified strains were quite stable and hence useful in 
epidemiological studies (31), similar to those used to monitor the 
spread of selected human pathogens (28). The purpose of this study 
was to survey the extent of variation of plasmid DNAs within a large 
number of X. campestris pathovars and to determine whether the plasmid 
content of strains based on restriction fragnent polymorphism and 
Southern hybridization could be used in differentiating the pathovars 
of X. campestris . 

Materials and Methods 
Bacterial Strains 

The X. campestris strains used in this study, their pathovar 
designations, geographic origin, and sources are listed in Table 4-1. 



Table 4-1. strains of X. carr.pestris used for plasmid analysis, 



29 



Bacterium 

(number of strains) 



Strain designation 
(geographic origin) 



Source^ 



Xanthononas campestris pathovar 



alfalfae (2) 

argemones (1) 
begoniae (1) 
campestris (4) 



carotae (3) 
citri (5+) 



cyamopsidis (5) 

dieffenbachiae (2) 
esculenti (1) 
glycines (3) 

hederae (3) 

holcicola (2) 

maculi f ol i igardeniae {2] 
malvacearum- cotton (32) 



malvacearum- hibiscus (8) 



mangiferaeindicae (1) 

pelargonii (1) 
phaseol i (13) 



KS (Kansas) ; 

FL (Florida) ; 

084-1052 (Florida); 

084-155 (Florida) ; 

XCl (Oklahoma); 

084-720,084-809, 

084-1318 (Florida) ; 

Gl (Idaho) ,G5,G7 (California) 

X59 (Brazil) ,X62 (Japan) , 

X69 (Argentina) , X70 (Brazil) 

Fll (Florida) ; 

13 D5; 

X002,X005,X0016, 

X0017 (Arizona); 

084-729,084-1373 (Florida); 

084-1093 (Florida) ; 

B-9-3 (Brazil) ,1717 (Africa), 

17915; 

084-1789,084-3928, 

251G (Florida); 

Xh66 (Kansas) ; 

XHl; 

084-6006,084-6166 (Florida); 

A,B,E,F,G,H (Oklahoma); 

HV25 (Upper Volta) , 

Chl,Ch2 (Chad), 

Su2,Su3 (Sudan); 

FL79 (Florida) ; 

D,M,N,0,U,V,W,X,Y,Z,TX84 (Texas) , 

I,Q,R,S,T (Oklahcma), 

C,J,K,L (Upper Volta) ; 

X10,X27,X52,X102, 

X103,X108 (Florida); 

083-4344, M84-11 (Florida); 

084-166 (Florida) ; 

084-190 (Florida) ; 

EKll,Xph25,Xpfll (Nebraska); 

Xpa,Xpll (Wisconsin); 

82-1,82-2 (Florida); 

LB-2,SC-3B (Nebraska); 

XP2 (New York) ; 

XP-JL (Kansas) ; 

XP-JF (Missouri) ,XP-DPI; 



D.L.Stuteville 

R.E. Stall 

DPI 

DPI 

this study 

DPI 

R.E. Stall 

E.L.Civerolo 

DPI 

C.I.Kado 

J.Mihail 

DPI 

DPI 

W.F.Fett 

DPI 

L.Claflin 
this study 
DPI 
M.Essenberg 



L.S.Bird 
DPI 



this study 



A.R.Chase 

DPI 

DPI 

DPI 

M. Schuster 

A-W.Saettler 

R.E. Stall 

A.K.Vidaver 

J.A.Laurence 

J. L. Leach 

this study 



30 



(Table 4-1 continued.) 



Bacterium 

(number of strains) 



pisi (1) 

poinsettiicola (1) 
pruni (3) 

translucens (2) 

vesicatoria (5) 

vignicola (7) 

vitians (3) 

zinniae (1) 

unknown (1) 

X. albilineans (1) 
X. fragariae (1) 



Strain designation 
(geographic origin) 



Source^ 



XPl (Japan) ; 

083-6248 (Florida); 

068-1008,084-1793 (Florida); 

82-1 (Florida); 

82-1 (Florida) ; 

XTl; 

E-3, 69-13, 71-21, 

82-8,82-23 (Florida); 

A81-331,C-1,CB5-1, 

Xvl9,SN2, 432, 82-38 (Georgia) 

084-2057,084-2848, 

084-4348 (Florida); 

084-1944 (Florida) ; 

G65 

Xalb (Florida) ; 
Xfra (Florida) ; 



KuGoto 

DPI 

DPI 

R.E. Stall 

R.E. Stall 

this study 

R.E. Stall 

R.D.Gitaitis 

DPI 
DPI 

this study 

M.J.Davis 
R.E. Stall 



^ DPI - Florida Department of Agricultural and Consumer Services 
Division of Plant Industry, Gainesville, Florida. ' 



Some of the stock cultures of x. c. malvacearum were mixtures of 
different strains, maintained and described as "races" for purposes of 
screening cotton host differentials and breeding lines for disease 
resistance. Described "races" of X. c. malvacearum may or may not be 
purified strains; as many as seven different strains have been derived 
from a single race 1 isolate (7) . All strains used in this study were 
repeatedly purified from single colonies and confirmed to be 
pathogenic on their designated hosts. Broth cultures of bacteria were 
grown in a peptone-glycerol medium (per L: 10.0 ml glycerol, 20.0 g 
peptone, and 1.5 g K2HPO4) . Bacterial strains were comionly stored at 
-80 C in the same medium containing 15% glycerol. 
Plasmid Extraction and Visualization 

Cultures were grown to mid- to late-logarithmic growth phase and 
extracted by either of two small-scale alkaline lysis extraction 
procedures (46,66). Extracted DNA was resuspended in TE (10 mM 
Tris(hydroxymethyl)aminomethane (Tris), 1 mM sodium 
ethylenediaminetetraacetate (NasEDTA) , and 20 ug/ml DNase-free 
pancreatic RNase; pH 7.6) and digested with either of two restriction 
endonucleases, EcoRl or BaMil, using manufacturer (Bethesda Research 
Laboratories, Gaithersburg, MD) specifications. Plasmid DNA fragments 
were separated by size using agarose gel electrophoresis (0.6% agarose 
(Sigma Type I:low EEO) , 2-5 V/cm) in Tris-acetate buffer (40 mM Tris, 
1 mM Ma2EDTA, adjusted to pH 7.6 with glacial acetic acid). Fragments 
were visualized by ultraviolet irradiation (302 nm) after staining 
agarose gels in ethidium bromide (0.5 ug/ml). Photographs were taken 
with Polaroid Type 55 (or Type 57) fiLm using a yellow filter (Tiffen 



31 



32 



no. 12). All restriction fragment size estimates were based on the 
relative mobilities of linear DNA fragments using lambda phage DNA 
digested with Hindlll as molecular size standards. Plasmid sizes were 
estimated by addition of the sized restriction digested plasmid DNA 
fragments. All plasmid experiments were repeated at least twice for 
each strain examined. 

Cloning of Plasmid Restriction Endonuclease Fragments 

Plasmid DNA isolation from X. campestris was by a modification of 
either of two alkaline lysis extraction procedures (46,68). The 
extracted plasmids were purified on CsCl-ethidium bromide gradients by 
centrifugation at 55,000 rpn in a Beclor.an VTi65.2 rotor for 17 hr at 
20 C. The purified plasmids were digested with the restriction enzyme 
EcoRI. The cloning vector, pUCDS (14), was digested with EcoRI, 
treated with calf intestinal alkaline phosphatase (Boehringer- 
Mannheim, Indianapolis, IN), and ligated to the EcoRI digested X. 
campestris plasmid fragments with T4 DNA ligase (Bethesda Research 
Laboratories). The ligation products were transformed into E. coli 
strain EDS 767, selecting transformed colonies on Luria-Bertani medium 
containing ampicillin (50 ug/ml) or kanamycin (30 ug/ml) . Selected 
colonies were analyzed for the vector containing desired cloned DNA 
fragnents. These general cloning procedures are outlined in Maniatis 
et al. (66) . 
DNA/DNA Hybridization 

Plasmid DNAs were transferred from agarose gels to nitrocellulose 

membranes by the method of Southern as described by Maniatis et al. 

(66) and hybridized against radioactively-labeled DNA probes. The DNA 



probes derived from plasmid DMA of X. campestris pathovars were either 
cloned restriction digested DNA fragments of plasmid DNA in the cosmid 
vector PUCD5, or of the complete X. campestris plasmid. DNA probes 
were labeled in vitro with use of a nick translation kit (Bethesda 
Research Laboratories) using 32p_deoxycytidine triphosphate and 
hybridized against DNA bound to nitrocellulose membranes. The 
m^ranes were pre-hybridized and hybridized in plastic bags at 68 C. 
The pre-hybridization and hybridization solutions were those described 
by Maniatis et al. Following hybridization, merrforanes were washed once 
m 2X SSC, 0.5% SDS and washed once in 2X SSC, 0.1% SDS at ambient 
temperature, and washed two times in O.lX SSC, 0.5% SDS at 68 C as 
described by Maniatis et al. (66) for stringent conditions. The 
membranes were then air dried and exposed at -80 C to X-ray film 
(Kodak X-Omat AR) using intensifier screens. similar methods were 
used to probe DNA transferred to nitrocellulose using a dot-blot 
manifold (Schleicher and Schuell inc., Keene, m) . All hybridization 
experiments were repeated at least once. 

Results 
Detection of Plasmid DNA 

Indigenous, cryptic plasmids were detected in all strains of the 
following X. campestris pathovars: cyamopsidis , dieffenbachiae, 
glycines, malvacearum (cotton), pelargonii , phaseoli , pruni, 
vesicatoria , vignicola, and vitians (Table 4-2). Plasmids were not 
detected in any strains of X. campestris pathovars alfalfae , 
argemones , begoniae , carotae , esculenti , holcicola , 
maculifoliiqardeniae, malvacearum (hibiscus), mangiferaeindicae , pisi. 



33 



34 



Table 4-2, Detection of plasmid DNA in strains of X. campestris. 



Bacterium 



No. of strains 
containing plasmids/ 
No. of strains tested 



X. campestris pathovar 

alfalfae 

argemones 

begoniae 

campestris 

carotae 

citri 

cyamopsidis 

dieffenbachiae 

esculenti 



glycines 
hederae 
hole i col a 



maculi f ol i igar deniae 

malvacearum - cotton 

malvacearum - hibiscus 

mangiferaeindicae 

pelargonii 

phaseoli 

pisi 

poinsettiicola 

pruni 

translucens 

vesicatoria 

vignicola 

vitians 

zinniae 

X. albilineans 

X . fragariae 



0/ 2 


0/ 1 


0/ 1 


2/ 4 


0/ 3 


17/44 


5/ 5 


2/ 2 


0/ 1 


3/ 3 


2/ 3 


0/ 2 


0/ 2 


32/32 


0/ 8 


0/ 1 


1/ 1 


13/13 


0/ 1 


0/ 1 


3/ 3 


0/ 2 


5/ 5 


7/ 7 


3/ 3 


0/ 1 


0/ 1 


0/ 1 



Pa th ogen i c i ty^ 



P 
I 

I 

R,l 
R 

P,R,I 

R 

I 

I 

P 

I 

R 

I 

P 

P 

I 

R 

P 
R 

I 

I 
R 
P 
P 
1 
I 
R 
I 



P - Pathogenicity of strains confirmed on appropriate host, conforms 
to current available information for particular pathovar. 

R = Received as named pathogen, appropriate host specificity and 
pathovar designation of strain assumed. 

LZ^^^f^^^^^^^ "" pathogen on host appropriate for designated pathovar; 
characterization of strain (s) incomplete. 



35 
poinsettlicola , translucens, and zinniae . Plasmids were found insome, 

but not all strains of x. campestris pathovars campestris and 
hederae. Similarly, plasmids were found in all type strains of x. c. 
pv. citri (A, B, and C types); this is in agreement with a previous 
placid study of this pathovar (13). However, not in all strains of x. 
CEnpestris isolated from leaf spots of citrus in Florida contained 
plaanids. 

It appeared that plasmid-containing strains of x. campestris 
carried from one to three plasmids based on electrophoresis of 
extracted plasmid DNA. For example, a majority of the 
plasmid-containing X. c. malvacearum strains contained only one 
Plaanid, but some carried two or more. when plasmid-containing 
strains of X. campestris were purified by repeated single colony 
isolation, the plasmid DNA content appeared to be stable. However, 
variation was present in bacterial stocks which were known to have 
been serially transferred in agar medium over a period of years. 
Restriction Endonuclease Profile s 

Plasmid profiles for x. cam^^stris were variable in over 60 
different strains tested. Plasmids were placed into classes based on 
the variability of restriction endonuclease (EcoRl) digestion profiles 
on agarose gels. Wnen a different restriction endonuclease (BamHI) 
was used, the plasmid profiles were placed into the same classes. In 
all cases, strains which belonged to the same plasmid class also 
belonged to the same pathovar. There was obvious variability within 
classes, but there also appeared to be conservation of some DNA 
fragments of identical sizes (Fig. 4-1). Undigested plasmids were not 



reliable for strain classxf ication because several strains had 
Pla^ids of apparently identical size, but they were clearly different 
after digesting the plasmid DNAs with restriction enzymes. By adding 
up the DNA fragment sizes yielded by restriction digests, plasmids in 
X. c^npestris were estimated as ranging from about 3 to 200 kb 
(kilobase pairs) in size. Estimation of some of these sizes were 

difficult for sane strains due to the presence of more than one 

plasmid. 

Plasmid profiles of strains of X. c. cyamopsidis , X. c. glycines , 
X. c. m alvacearum , X. c. ^^aseoli, and X. c. vignicola were compared. 
Restriction fragment length polymorphism was evident within each of 
these pathovars. Although more than one plasmid was present in some 
Of these strains, a subset of restriction fragments of similar length 
and overall pattern appeared to be consistent for strains within a 
given pathovar (Figure 4-2) . 

Four highly poli^orphic plasmid variants were found in 17 out of 
44 Florida strains tested, but Southern hybridization revealed no 
hanology between sa.e of the plasmids (31). Furthermore, there were 
no similarities in plasmid digestion patterns between X. c. pv. citri 
ti^ A, t,pe B and any of the Florida citrus leaf spot strains that 
carried plasmids (Figure 4-3). These Florida isolates are presumed to 
be X. c. citri, but they grow well and also cause disease s^^ptcms on 
kidney bean and alfalfa, thus making their pathovar status 
questionable (34) . 



36 



37 



A BCDE FGH I JKLMNOPQR 




DNAs from strains of Xanthononas campestris pv. 



Figure 4-1. Plasmid 

malvacearum ( Xcm ) digested with restriction endonucleases EcoRl (lanes 
B-I) and BamH I (lanes J-Q) . Lanes shown above contain: A and R, lambda 
Hindlll; B and J; Xon J; C and K, Xcm N; D and L, Xcm H; E and M, Xon 
V; F and N, Xon Z; G and 0, Xon Q; H and P, Xori X; and I and Q, Xon D. 



38 





1 


2 3 4 5 6 7 8 


111111 1 111222 
901234567 8 9012 


22 2 222233333333334444444 
345678901234567890123456 










I r 




30- 




— — 


25- 


— 


20- 




is- 




— , 














m- 


_ 
















s' 




















1 1 1 1 1 1 1 1 1 






0- 

kb 





Figure 4-2. Graphic representation of plasmid restriction fragment 
profiles for Xanthcmonas campestris pvs. cyamopsidis (Xcc; lane 1) , 
vignicola ( Xcv ; lanes 2-8), phased i ( Xcp ; lanes 9-17) , "5iaseoli var. 
fuscans ( Xcpf ; lanes 18-19) , 

lanes 23-46) 



maivacearum (Xan; 



glycines (Xcg; lanes 20-22) , and 
digested with EcoR I . Lanes shown above 
contain 1, Xcc 13D5; 2, Xcv SN2; 3, Xcv A81-331; 4, Xcv C-1; 5, Xcv 
CB5-1; 6, Xcv Xvl9; 7, Xcv 82-38; 1 



82-2; 



Xcv 432; 9, Xcp EKll; 10, Xcp 

11, Xcp Xpa; 12, Xcp Xpll; 13, Xcp XP2; 14, Xcp XP-JF; 15, Xcp 
XP-DPI; 16, Xcp Xph25; 17, Xcp 82-1; 18, Xcpf Xpfll; 19, Xcpf SC-3b7 
20, Xcg B-9-3; 21, Xcg 17915; 22, Xcg 1717; 23, Xon J; 24, Xon L; 25, 
Xon C; 26, Xoti 0; 27, Xon K; 28, Xcm M; 29, Xciti A; 30, Xcm B; 31, Xon 
E; 32, Xcm F; 33, 



Xcm G; 34, 



Xon H; 35, Xon S; 36, Xor W; 37, Xon I; 

38, XoTi V; 39, Xori D; 40, Xon M; 41, Xcm X; 42, Xctti U; 43, Xon Q; 44, 
XoTi R; 45, Xon Y; and 46, Xon Z. Calculated DNA fragment sizes are 
represented here by a linear scale whereas migration of DNA fragments 
under electrophoresis conditions approximates a logarithmic scale 
v/hich is inversely proportional to the molecular weight of the DNA 
fragment. 



39 




Figure 4-3. Plasmid DNAs frcm strains of X. campestris pv. citri 
aigested with restriction endonuclease EcoRI. Lane A, strain"~FlT 
(Florida ti^); lane B, strain X59 (A type); lane C; strain X62 (A 
type); and lane D, strain X69 (B type) , sizes shown in kilobase pairs 



mh/rUA HvbridjyflHon 

Initial plasmid comparisons were done on strains of X o 
-i-ea™. *ole purified pias^id DMA frc„ x. c. jalvacea™ s^^rain 
X, which contains only one pla^xd, „as hybridized against EcoRI 
digested placid DNAs of other strains of the s^e pathovar (not 
shown). This initial co:„parison demonstrated that the 
Plas^ids.although differing slightly in digestion patterns, were quite 
homologous as the radiolabeled placid hybridized to almost all EcoRI 
fragments of the otlier strains. 

A 4.5 kb ECORI placid fra^ent of x. c. malvacear^ strain N was 
Cloned into the vector pUCD5 and used as a hybridization probe against 
Plasmid DNA fro, other strains of the s^e pathovar fflg. 4-4) . This 
Placid DMA fragment hybridized to plasmids fron all but one strain of 
X. c. m alvacearnm . This probe hybridized to EcoKl fragments of 4.5 kb 
size (lanes B-J, M-N, Fig. 4-4) in several other strains of x. c. 
KiSaarS. Additionally, the probe hybridized to more than one of 

the EcoRl plasmid fragments in the^P c:+-v=.'„^ 

^fc^icfa m tftese strains, suggesting that sane 

sequences on the cloned mA are reflated In other parts of the plasmid 
DMA. The paCDS vector alone did not hybridize to any x. c. 
Eii2«asa Plasmid fragments. Because puCD5 is a cosnid vector and 
contains the cos site of la.TMa phage DNA, hybridization of the vector 
to the corresponding DNA fragment of the molecular weight mar)=er 
containing the cos site was observed. Placid DNA from X. c. 
ESiHHS™ strain Su2 did not hybridize to the 4.5 kb probe, and did 
not hybridize to any other plasmid fragments fro, X. c. malvacearum 
strain N. Another X. c. malvacearnm strain ,ch2) , which did not have 



40 



41 



ABCDEFGHIJKLMN 




Figure 4-4. Plaanid DNAs from strains of Xanthomonas campestris pv. 
malvacearum ( Xcm ) digested with restriction endonuclease EcoRI (left) 
and autoradiograph of plasmid DNAs probed with a clone containing a 
4.5 kb Eco RI plasmid fragment from Xcm strain N (right) . Lanes shown 
above contain: A, lambda Hind lll B, Xcm H; C, Xcm W; D, Xcm V; E, Xcm 
X; F, Xon Q; G, Xon Y; H, Xon L; I, Xoti N; J, Xcm HV25; K, Xan Ch2; L, 
Xcm Su2; M, Xcm FL79; and N, Xcm TX84. 



42 



a 4.5 kb EcoRI plasmid DNA fragment, did have two other fragments (ca. 
8 and 10 kb) which hybridized strongly to the 4.5 kb probe. In two 
other X. c. malvacearum strains (FL79 and TX84) , which appeared to 
have multiple plasmids, the hybridization signal was V7eak for the 4.5 
kb fragments as compared to larger EcoR I fragments (ca. 23 kb) . 
Similar hybridization studies were done with cloned plasmid fragments 
from other pathovars. When the plasmid fragments selected were 
smaller, they were much more specific. Two different probes were 
constructed from 2.0 kb and 2.3 kb EcoR I plasmid fragments derived 
from X. £_. phaseoli (strain XP2) . When hybridized against X. c. 
cyamopsidis , X. c. glycines , X. c. phaseoli , and X. c. vignicola, each 
of these probes only hybridized to X. c_. phaseoli (Fig. 4-5) . The 2.3 
kb plasmid probe hybridized to similar sized fragments in other X. c. 
phaseoli strains, including X. c. phaseoli var. fuscans (strains SC3-B 
(not shown) and Xpfll) . These X. c_. phaseoli var. fuscans strains 
differ from typical X. c. phaseoli strains in that they produce an 
extracellular dark brown, melanin-like pigment in culture; otherwise, 
they are considered similar. However, the 2.0 kb plasmid probe did 
not hybridize to the X. c. phaseol i var. fuscans strains, which did 
not have the corresponding 2.0 kb EcoR I fragment in their plasmid 
profile (lane C, Fig. 4-5, and strain SC3-B (not shown)). Repeated 
hybridizations with Southern transfers containing these strains had 
the same results. 
Dot-blot Hybridization 

-»■ 

DNA probes were also hybridized against total DNA of other X. 
campestris pathovars fixed onto a nitrocellulose membrane by use of a 
dot-blot manifold apparatus. Radiolabeled total plasmid DNA from 



43 



A B C D E F G H I J K L ABC DEFGHIJ KL ABCDE FGHIJKL 







** 



% .• 



..*— •• 



Figure 4-5. Plaanid DNAs from strains of Xanthomonas campestris pvs. 
phaseoli (Xcp; lanes B, I>-I) , phaseoli var. fuscans ( Xcpf ; lane C) , 
and cyamopsidis (Xcc; lane K) digested with restriction endonuclease 
Eco RI (left) and autoradiographs of plasmid DNAs probed with a clones 
containing a 2.0 kb EcoRI fragment (center) and a 2.3 kb Eco RI plasmid 
fragment (right) from Xcp strain XP2. Lane J contains chromosomal DNA 
of X. campestris pv. alfalfae ( Xca ) digested v/ith Eco RI. Lanes shown 
above contains: A and L, lambda Hindi II; B, Xcp Xph25; C, Xcpf Xpfll; 
D, Xcp EKll; E, Xcp 82-1; F, Xcp Xpa; G, Xc G65; H, Xcp Xpll; I, Xcp 
XP-JF; J, Xca FL; and K, Xcc 13D5. Xc G65 is a strain isolated from 
bean which was determined to be nonpathogenic and contains no plasmid. 



44 



Table 4-3. Hybridization of radiolabeled plasnid probes to total DNA of 
pathovars of XanthoTionas campestris and one other Xanthononas species. 



Bacterium (No. tested) 



Probes^ 



N80 N4.5 V2.3 P2.3 



X. campestris pv. 

alfalfae (1) 
argemones (1) 
begoniae (1) 
campestris (1) 
carotae (1) 
citri (3) 
cyamopsidis (1) 
dieffenbachiae 
esculenti (1) 
glycines (1) 
hederae (1) 
hole i col a (1) 



(1) 



maculifoliigardeniae (1) 



vitians (1) 
zinniae (1) 



X. albilineans (1) 



+ 
+ 
+ 



+ 
+ 



+/_b +/_b +/_b 
+ + - 



malvacearum - cotton (6) 


+/-C 


+/-^ 


+ 


malvacearum - hibiscus (2) 


- 




— 


mangiferaeindicae 


(1) 


- 


— 


— 


phaseoli (1) 




+ 


+ 


— 


poinsettiicola (1) 




— 


— 


_ 


pruni (1) 




+ 


+ 


— 


translucens (1) 




+ 


— 


_ 


vesicatoria (1) 




+ 


+ 


+ 


vignicola (1) 




+ 


+ 


+ 



^ + = hybridization observed, - = no hybridization observed. 
N80 = plasmid DNA derived from X. c. pv. malvacearum strain N 
(about 80 kb) ; N4.5 = cloned EcoR I plasmid fragment (4.5 kb) from 
strain N; V2.3 = cloned EcoRI plasmid fragment (2.3 kb) from X. c. 
pv. vignicola strain SN2; P2.3 = cloned EcoRI plasmid fragment 
(2.3 kb) from X. £. pv. phaseoli strain XP2 

" Strain FL-11 was negative, only plasmid DNA used. 



Strain S2 was negative. 



45 



X. c. malvacearurn strain N, which carries two plasmids, hybridized to 
EffJA of thirteen out of twenty-three X. campestris pathovars tested 
(Table 4-3). Of these thirteen, seven cross-hybridized to the 4.5 kb 
subcloned fragment of X. c. malvacearum strain N. Plasmids were not 
detected in seme of the pathovars which hybridized to the probe. A 
2.3 kb subcloned plasmid fragment from X. c. vignicola hybridized to 
only six of the 23 pathovars tested. Plasmids were present in all six 
of those pathovars which hybridized to the probe. A 2.3 kb cloned 
plasmid fragment of X. c. phaseoli hybridized strongly to other 
strains of the same pathovar, and weakly to 2 of 3 different X. c. 
citri strains tested. The X. c. phaseoli probe appears to have 
hybridized to chromosomal DNA of X. c. citri in this case, as the 
probe did not hybridize against Southern transfers of EcoRI digested 
plasmid fragments of X. c. citri strains (not shown) . 

Discussion 
It is suggested that there is extensive conservation of plasmid 
DNA sequences (as represented by conserved restriction fragments) 
within, but not usually among, pathovars of X. campestris . Plasmid DNA 
fragment patterns will be identical if there is no rearrangement of 
the DNA sequence at the restriction enzyme recognition site (a six 
base pair sequence for Eco RI and Bam HI ) , if no new restriction 
fragments are created within the fragment, and if there are no major 
additions or deletions causing a change in fragment size. Given these 
possibilities, it was surprising to find so little restriction 
fragment length polymorphism of plasmids within a pathovar, especially 
v;hen strains obtained from different continents were compared. 



46 



Southern hybridization analyses confirmed that plaanid DNA fragments 
of similar size were in fact highly homologous. For example, X. c. 
malvacearum strain N, isolated in North America, had a restriction 
digest pattern identical to one of the African strains (K) (Fig. 4-2) . 
All North American strains except strain N form a distinctive pattern 
subgroup and the African strains form a somewhat different subgroup. 
It is possible that either all the African strains are derived fron 
strain N, or that strain N was introduced to the North America from 
Africa. The latter possibility seems more likely, since the African 
strains were isolated from more than one location in Africa. The 
conservation of overall restriction fragment profiles of plasmids from 
geographically isolated populations and the extent of hcmology seen in 
Southern hybridizations strongly suggests that plasmid sequences are 
both highly conserved and stable. 

Conversely, it was surprising to find so much polymorphism of 
plasnids between pathovars. Homology among plasmids in different 
strains of one pathovar of Pseudcmonas syringae (pv. glycinea) has 
been reported (15) . However, identical plasmid profiles were found 
present in more than one pathovar of P. syringae (77) . In the present 
study, similar plasmid profiles were not found in more than one 
pathovar. In addition, cloned plasmid fragments were identified v/hich 
failed to hybridize to plasmids of different pathovars. For example, 
a cloned 2.3 kb plasmid fragment of X. c. phaseoli , v^ich hybridized 
to all strains of that pathovar, failed to hybridize to total DNAs of 
22 other pathovars. This DNA fragment is highly conserved and is 
apparently pathovar-specif ic, with one exception (Table 4-3) . 



47 



Plasmid homology between X. c. phaseoli and X. c. phaseoli var. 
fuscans was revealed by hybridization that was not apparent by 
restriction digest patterns. There were distinct differences in 
plasmid digestion patterns of X, c. phaseoli var, fuscans in 
conparison to typical X. c. phaseoli strains. The fact that both X. 
c_. phaseoli and X. c_. phaseol i var. fuscans have an identical host 
range suggests that the homologous plasmid DNA regions that are 
conserved within seme pathovars may encode host range specificity 
functions. Such functions have been described on plaanids within the 
Rhizobiaceae (62,64) . This is a testable hypothesis that needs 
further experimental support. 

In X. c. malvacearum , plasmids were found in all 32 strains 
isolated from cotton, and no plasmids were found in the 8 strains 
isolated from hibiscus. Cotton and hibiscus each belong to the same 
plant family (Malvaceae) , hence the X. c. malvacearum designation. 
Atypical symptons of cotton blight could be artificially produced in 
cotton using syringe inoculations with hibiscus strains, and in 
hibiscus using cotton strains. The ability of these strains to be 
pathogenic on both hosts under natural conditions has not been clearly 
established. Plasmid DNAs from cotton strains cross-hybridized with 
one another in Southern analyses, indicating extensive homology was 
evident between the plasmids. As with X. c. phaseoli , the cotton 
strains of X. c_. malvacearum appeared to carry highly conserved 
plasmid DNA sequences (similar in restriction digest sizes and by 
Southern analyses) which are unique to strains which have a host range 
on cotton. Plasmid DNA from the cotton strain (N) , when radiolabeled. 



48 



did not hybridize to chromosomal DNA derived fron the hibiscus strain. 
Based on our limited pathogenicity tests and on the absence of plasmid 
DNA sequences in the hibiscus strains, there may be justification to 
differentiate the cotton and hibiscus strains into different 
pathovars. 

Although some cross-hybridization between plasmids of strains 
fron different pathovars was detected by dot-blot analyses (Table 4- 
3), strains were readily differentiated by restriction digest profiles 
of the plasmid DNAs and by hybridization with selected DNA probes to 
identifiable restriction digested DNA fragments. Cross-hybridization 
may be the result of repetitive DNA sequences, insertion elements, or 
of conserved DNA sequences that are important for the stable 
maintenance of plasmids in X. campestris . Examples of plasmid 
functions which might be conserved are those involved with 
replication, incompatibility, or other host dependent factors. In 
seme instances plasmid DNAs hybridized to total DNA of pathovars in 
which no plasmids were detected (Tables 4-2 and 4-3) . It is suggested 
that some sequences encoded on plasmids in some pathovars are located 
on chromosomes in other pathovars. It is also possible that those 
plasmids may integrate into the bacterial chromosome and excise again, 
in a manner similar to those in P. syringae pv. phaseolicola (79,92). 

With interest in developing rapid diagnostic methods to identify 
bacterial pathogens, it is possible that the combined usage of plasmid 
restriction digest profiles and of plasmid DNA probes may be 
sufficient for the identification of seme pathovars of X. campestris . 
This would require that a plasmid be stably associated with a given 



49 



pathovar, that plasnid profiles for specific pathovars were known, and 
that an appropriate DNA probe consisting of conserved and unique DNA 
sequences were available. Of the few DNA probes constructed, it was 
apparent that plasmid DNA sequences were highly conserved within the 
pathovars studied. In seme instances the DNA probes may prove 
sufficient for pathovar identification, provided they are extensively 
tested. This apparent stability of the plasmids provides a natural 
genetic marker that can be strain specific and perhaps useful in 
epidemiological investigations. In addition to aiding the 
identification for some pathovars of X. campestris , these observations 
may have taxoncmic significance to differentiate these pathogens. 



CHAPTER FIVE 

AEIE RACE-SPECIFIC AVIRULENCE GENES ENCODED BY PLASMIDS 

OF XANTHOMONAS. CAMPESTRIS PV. MALVACEARUM? 



Introduction 

Bacterial blight of cotton, caused by the pathogen X. c. pv. 

malvacearum (Shiith) Dye, is a destructive disease of cotton, which 

occurs throughout the cotton-producing areas of the world. The 

development of cotton cultivars resistant to bacterial blight is the 

primary means by which this pathogen is controlled. At least 16 major 

genes have been identified which condition resistance to bacterial 

blight (8) . A gene-for-gene pattern of interaction between the plant 

resistance genes and bacterial avirulence genes has been demonstrated 

in this host-parasite interaction (33) . In screening for resistance 

to bacterial blight in breeding lines, races of the pathogen have been 

maintained as mixed cultures, originally derived from infected leaf 

samples. These mixed cultures have been mass transferred on agar 

media to prevent the appearance of new race phenotypes. The stability 

of the race phenotypes of these mixed cultures is periodically checked 

by inoculating the host differentials. 

Plasmids have been detected in several phytopathogenic 
prokaryotes and have been associated with functions such as 
pathogenicity (5), resistance to antibiotics (3,88), avirulence (88), 
and host range (64) . VJith the recent discovery of plasmids in several 

SO 



51 



pathovars in X. campestris (13,16,30,39,46,52,53,80), it is of 
interest to understand the function of these autonanously replicating, 
covalently closed circular DNAs of the bacterial genome. The purpose 
of this investigation was to analyze the plasmids in X. c. pv. 
malvacearum and determine if they influenced race-specificity in 
cotton. 

Materials and Methods 
Bacterial Strains and Host Plants 

The strains of X. c. pv. malvacearum used for this investigation 
are listed in Table 5-1. All of the strains were from single colony 
isolation of either mixed, or pure cultures. The cotton differentials 
used included those recommended for differentiating the races of X. c. 
pv. malvacearum (45) ; also included were six congenic cotton lines of 
Gossypium hirsutum cv. Acala 44, which contained single known 
resistance genes (B genes) to X. c. pv. malvacearum (7) . The congenic 
Acala series included cotton with resistance genes B2, B3, B5, Bg, B7, 
and B^. The other host differentials were cvs. Acala 44, 1-lOB, 101- 
102B, 20-3, Msbane Bl, Stoneville 20, 2B-S9, and Gregg. 
Plant Inoculations 

Pathogenicity tests were conducted by pressure infiltration of 
abaxial leaf surfaces with bacterial suspensions using a blunt end of 
a syringe, and incubating the plants in 30 C growth chambers until 
symptom were expressed. Symptoms development was monitored over a two 
week period. 
Plasmid Analysis 

Cultures were grown to mid- to late-logarithmic grov7th phase and 
extracted by either of two small-scale alkaline lysis extraction 



Table 5-1. Strains of X. campestris 
in race- specificity investigation. 



pv. malvacearum from cotton used 



Strain 



Race 



Location 



Source 



K 


HVlb 


Upper Volta 


L.S.Bird 


N 


2a 


Texas 


L.S.Bird 


J 


HV3 


Upper Volta 


L.S.Bird 


C 


HVla 


Upper Volta 


L.S.Bird 





2b 


Texas 


L.S.Bird 


L 


HV7 


Upper Volta 


L.S.Bird 


A 


2hS^ 


Oklahoma 


K. Essenberg 


B 


2aS^ 


Oklahona 


M. Essenberg 


E 


4-1 


Oklahotia 


M. Essenberg 


F 


laS^^ 


Oklahcma 


M. Essenberg 


G 


IbS^^ 


Oklahoma 


M. Essenberg 


S 


3-22 


Oklahoma 


M. Essenberg 


H 


4-2 


Oklahoma 


M. Essenberg 


W 


6 


Texas 


L.S.Bird 


I 


3 


Oklahana 


M. Essenberg 


V 


3 


Texas 


L.S.Bird 


D 


2 


Texas 


L.S.Bird 


M 


15 


Texas 


L.S.Bird 


X 


18 


Texas 


L.S.Bird 


u 


1 


Texas 


L.S.Bird 


Q 


la 


OklahoTia 


M. Essenberg 


R 


lb 


Oklahoma 


M. Essenberg 


Z 


7b 


Texas 


L.S.Bird 


Y 


7a 


Texas 


L.S.Bird 


T 


18 


Oklahcsna 


M. Essenberg 


Chi 


- 


Chad 


L.S.Bird 


Ch2 


- 


Chad 


■L.S.Bird 


HV25 


- 


Upper Volta 


L.S.Bird 


Su2 


- 


Sudan 


L.S.Bird 


Su3 


- 


Sudan 


L.S.Bird 


FL79 


16 


Florida 


this study 



53 

procedures (46,66) as described in Chapter Four. Plasmids were 
digested with either Eco RI or Bam HI and DNA fragment patterns 
visualized by agarose gel electrophoresis. All plasmid experiments 
were repeated at least twice for each strain examined. Additionally, 
EcoRI plasmid fragments of X. c. pv. malvacearum strains were cloned 
into the vector pUCDS (14) and maintained in E. coli ED8767 as 
described in Chapter Four„ These cloned fragments were used for 
subsequent plasmid studies. 
Plasmid Curing 

To determine if plasmid encoded functions were involved in 
pathogenicity, attempts were made to expel, or cure, plasmids from the 
wild-type strains by including ethidium-bromide or sodium lauryl 
sulfate (SDS) in the broth culture. Cultures were grown in peptone- 
glycerol broth for two 48 hr cycles, being adjusted to an optical 
density (ODggonm) °^ 0.03 at the beginning of each cycle. A dilution 
series frcxn the broth cultures were plated onto peptone-glycerol agar 
and single colonies were selected for inoculation on cotton host 
differentials. 
Race-Specificity Genes 

Cloned genes encoding race-specific avirulence in X. c. pv. 
malvacearam have been identified and isolated (33) , and it was of 
interest to determine if these genes were plasmid-borne . Using E. 
'^PlJ- ED8767 containing the cloned plasmid fragments in pUCD5 as 
donors, a tri-parental mating scheme was used to move the cloning 
vector into strains of X. c. pv. malvacearum by conjugation (22,69). 
Into the recipient, X. c. pv. malvacearum. strain N (virulent strain) , 
were mated cloned plasmid fragments (in pUCD5) from X. c. pv. 



54 

malvacearum strain H (avirulent strain). As a control, plasmids 
fragments from strain N were also moved into strain H. The helper 
plasmid used for mobilization of the plasmids from E. coli into X- 
campestris were pSa322 or pRK2013 (22,93). The donor and helper 
strains were grown overnight at 37 C in LB broth with appropriate 
antibiotic selection pressure. The antibiotic was washed fran the 
cells prior to mating by centrifuging and resuspending the cells in 
0.7% NaCl. The X. campestris recipients were grown overnight at 30 C 
in MOPS minimal medium broth (69) . Donors and helpers at a 
concentration of 10^ cells and recipients at 10^ cells were mixed in a 
small volume onto sterile 25 mm cellulose-acetate filter membranes and 
placed onto peptone-glycerol medium for 6 hours. The membranes were 
then transferred to MOPS minimal medium with 25 ug/ml kanamycin and 
incubated for 72 hours at 30 C. Bacterial growth from filters was 
diluted in 0.7% NaCl and plated for single colony isolation on MOPS 
minimal medium with 25 ug/ml kanamycin. Colonies were selected and 
inoculated on the Acala cotton differentials containing known 
resistance genes. 
Plasmid Origin of Replication and Mobilization 

An assay was included to determine if X. c. pv. malvacearum 
contained self-mobilizing plasmids. The E. coli clones containing 
cloned plasmid DNA fragments from x. c. pv. malvacearum strain X were 
mated into spectinomycin resistant X. c. pv. malvacearum strain N in 
the presence, or absence, of the E. coli helper plasmid strain with 
PRK2013. Transconjugants were selected on MOPS minimal medium with 
kanamycin (25 ug/ml) , or directly on POPS (33) containing kanamycin 
(30 ug/ml) and spectinomycin (100 ug/ml) . 



55 



Results 
I Plasmids and Race-Specificity 

All of the strains of X. c. pv. malvacearum tested contained at 

least one plaanid. The plasmid sizes ranged fron about 50 to 100 kb. 

A majority of the American strains of X. c. pv. malvacearum contained 

J 4.8, 4.7, and 4.5 kb EcoRI DNA fragments which appeared to be 

conserved (Fig. 4-2). The entire 4.5 kb EcoRI DNA fragment appeared 

i; 

to be conserved in all but 1 of the 25 strains tested, regardless of 
geographic origin. 

Conparisons were made between the physical restriction fragment 
patterns (Fig. 4-2), the original race designations, and reactions on 
the 14 cotton differentials tested (Table 5-2) . The strains of X. c. 
pv. malvacearum were grouped into seven plasmid types based on the 
restriction profiles and the conservation of related restriction 
fragments. In the plasmid groupings I, in, and IV, a limited level 
of polymorphism in the restriction fragment profiles for each group 
were evident, but a majority of the fragments appeared to be 
conserved. In plasmid groupings II, V, VI, and VII, no polymorphisms 
were observed. 

The strains which contained plasmids of group I appeared to be 
pathogenic on a majority of the cotton host differentials. X. c. pv. 
malvacearum strain J was an exception for it was pathogenic to only 4 
of the cotton differentials, whereas the other group I strains were 
pathogenic on at least 11 of the cotton lines. Analysis of the EcoRI 
restriction fragment pattern of strain J revealed it was missing a 4.5 
kb fragment, and possibly had rearrangements which resulted in a shift 
of two other EcoRI fragments. With few exceptions, the 4.5 kb EcoRI 



56 



Table 5-2. Plasmid groupings and cotton plant reactions to strains of X. 
campestris pv. malvacearum. ^' 



B 



C 



K HVlb 
N 2a 
J HV3 
C HVla 
2b 
L HV7 
A 2bS^ 
B 2aS^ 
4-1 
laS^ 
Ibsr 
3-22 



H 4-2 



W 
I 
V 
D 



Y 

T 



A 
B 

C 

D 
E 
F 
G 
H 
I 
J 
K 
L 
M 

N 





6 
3 
3 
2 



M 15 

X 18 

U 1 

Q la 

R lb 

Z 7b 



7a 
18 



I 

I 

la 

lb 

lb 

Ic 

II 

II 

II 

II 

II 

II 

II 

Ilia 

Ilia 

lllb 

IVa 

IVb 

V 

V 

VI 

VI 

VII 

VII 



U 



80 
84 
59 
59 
59 
57 
82 
83, 
83. 
83. 
83. 
83. 
80. 
88. 
88. 
87. 
94. 
93. 
57. 
62. 
100. 
91. 
93. 
98. 



,7 

.8 

.5 

.0 

.0 

.9 

.2 

.3 

.3 

.3 

.3 

.3 

.4 

.7 

.4 

.0 

.2 

.2 

.5 

9 

4 

6 







2 
2 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

1 

2 

2 

1 

1 

1 



Inoculation reaction of cottonV 

+ + + + + + + + + + + + + + 14 14 

+ + + + V + V + + + + + + + 12 14 

+ ---- + v--- + -__ 34 

+ + + VVV + + VV + VV+ 7 14 

+ -V-V + V + VV + VVV 4 12 

+ ---V + + VVV + VVV 4 11 

+ - + -- + -----V-- 34 

+ - + -- + ---_-v-- 3 

+ -----v------_ 1 

+ - + -- + -----V-- 3 

+ - + -- + -----V-- 3 

+ ---------____ 1 

+ ------------- 1 

+ --V ------____ ]_ 

+ ---V-V------V 1 

+ ---VVV---V--- 1 

+ _ + __ + ____^ + __ 4 

+ - + -- + V----V-- 3 

+ -V-VV + ---VVVV 2 

+ -V-V + VV--VVVV 2 10 

+ ---------____ 11 

+ --VV-----V--- 14 

+ -V- + VV V--VVVV 2 10 

+ ---v-v------_ 13 

+ - + -- + ----- + -- 44 



4 
2 
4 
4 
1 
1 
2 
4 
5 
5 
5 
9 



or race 



= Strain designation. 

= Original strain, 

designation. 

= Eco RI plasmid group. 

= Kilobase pairs of plasmid DNk, 

= Number of plasmids. 

= Cultivar Acala 44. 

= Cultivar Acala B2. 

= Cultivar Acala B3. 

= Cultivar Acala B5. 

= Cultivar Acala Bg. 

= Cultivar Acala By. 

= Cultivar Acala B^. 



(B2 



Cultivar 1-lOB B 



'IN 



(+ polygenes) . 

= Cultivar 101-102B (B2B3 

+ polygenes) , 

= Cultivar 20-3 (Bj,j + polygenes) . 



P = Cultivar Mebane Bl 

polygenes) . 
Q = Cultivar Stoneville 20 (B7 

+ polygenes) . 
R = Cultivar Polygenes (2B-S9) 
S = Cultivar Gregg (resistance 

genes unknovm) . 
T =: Sum of cultivars which gave 4 

out of 4 compatible reactions 

(+). 
U = Total of cultivars which gave 

at least 1 compatible reaction 

(+). 
V = Plant reactions: +, compatible 

reaction; -, incompatible 

reaction; v, reaction variable 

over 4 replications. 



57 
fragment appeared to be conserved in all group I strains of X. c. pv. 

malvacearum. The group I strains K and N each contained two plasmids 
and also appeared to be the most virulent of all the strains tested in 
any group, each being pathogenic on all of the cotton differentials 
tested (Table 5-2). Strain K was isolated in Upper Volta, Africa and 
strain N was isolated from Texas. Although the plasmids in group I 
were highly conserved and apparently widespread in the world, there 
was no correlation between the presence of these plasmids with race 
specificity. 

The strains which were classified into plasmid group II could be 
divided into two basically different pathogenicity patterns on the 14 
cotton differentials. The strains which were derived from X. c. pv. 
malvacearum races 1 or 2, and strains derived from races 3 and 4 
appeared to have related pathogenicity patterns. The plasmid EcoRI 
restriction fragments patterns for all seven of these strains appeared 
identical. Similar variability was observed in the pathogenicity 
patterns for the other strains placed within groups ill, iv, V, VI, 
and VII. Again, no correlation between the pathogenicity pattern and 
the plasmid group vras apparent. The original race designations of the 
strain mixtures of X. c. pv. malvaceargm did not correspond to the 
race phenotypes exhibited by tlie strains purified from the mixtures. 
For instance, of 5 strains derived from nominal race 1 (F, G, Q, R, 
and U) , there were basically three different pathogenicity patterns 
observed, none of which corresponded to the reported pattern for race 
1. The expected race 1 reaction would have been pathogenic on Acala 
44 and Stoneville 2B-S9 (45) . 



58 

Plasmid Curing 

In preliminary tests with the curing agents, it was determined that 
0.3% SDS and 20 ug/ml ethidium brcmide were the concentrations which 
allowed growth of x. campestris . Strains did not grow at 
concentrations of 0.5% SDS or 30 ug/ml ethidium bromide. The 
bacterial colonies isolated from ethidium bromide treatments appeared 
phenotypically identical to the wild-type strains. Several colorless 
colonies were observed with the SDS treatment; however, over a 48 hr 
period, color appeared to be restored. Of 11 strains selected from 
the SDS treatment, all were found to be pathogenic on Acala 44. From 
plasmid analyses it appeared that 7 of 11 strains were missing the 
indigenous plasmids (Fig. 5-1) . 
Race-Specificity Genes 

Several clones containing cloned X. campestris plasmid fragments in 
the vector pUCD5 were obtained and transformed into E. coli ED8767. 

Based upon restriction analyses, only 3 of the 8 EcoRI fragments of 
strain H, and 6 of 9 EcoRl fragments of strain N were cloned. These 

clones carried 30 kb (36%) of the 84 kb strain H plasmid, and 70 kb 
(88%) of 85 kb strain N plasmids. Each of the E. coli clones were 

mated into X. c pv. malvacearum. The transconjugants were then 

selected and inoculated onto the 14 cotton differentials. The results 

of the inoculations are seen in Table 5-3. 

None of the clones derived from plasmid DNA of x. c. pv. 

malvacearum changed the expected disease reactions of X. c. pv. 

malvacearum strain H. One clone of X. c. pv. malvacearim strain H, 

originally thought to carry a large plasmid insert, did carry an 



59 



Table 5-3. Pathogenicity of X. campestris pv. malvacearum (Xcm) 
t ran scon jugantsa on cotton host differentials. 

Plant reactions^ 
Strains ^^ala 

~^ % ^ % B^^ ^ % B^ % %^ 101 Gregg BJ5 

wild-type: 

N + + + + + + + + + + + + + 

H +---------___ 

Plaanid fragtimts of Jtoi H in ^ N: 

pM^m + + + + + + + + + + + + + 

E«E?HB + + + + + + + + + + + + + 

Plasnid fragtsTts of ^ N in ^ H: 

E0E?HsI1 +--------____ 

f«E5r-N3 +-------_____ 

pJFMM +-------_____ 

FUE?ir-N5 +-------_____ 

ChranosaiHl fragrH* of }toi H in )&ti N: 

FtESr-HL + + + + + + + + _ + + + ^ 

a Cloned EcoRI fragment sizes within each clone are: pUFA-Hl, 4.9 kb; 
PUFA-H2, 19.2 and 4.6 kb; pUFA-H5, 22.4 and 4.0 kb; pUFA-Nl, 2.3 kb; 
PUFA-N3, 14.5; pUFA-N4, 23.5 and 13.4 kb; pUFA-N5, 15.6 kb; pUFA-N9', 
13.4 and 9.6 kb. 

+ = compatible (pathogenic) and - = incompatible (hypersensitive) 
plant reactions. 



60 




Figure 5-1. Plasmid curing of X. campestris pv. malvacear'um with SDS 
treatment. Lanes A-L are plasmid extractions of strains which were 
exposed to 0.3% SDS. Lanes A, E, I, and J appear to contain the 
indigenous plasmid. Lanes B-D, F-H, and K-L appear to be missing the 
indigenous plasmid. 



61 



avirulence gene. Transconjugants of this clone, pUFA-Hl, in strain N 
caused incompatibility on the Acala % line (Table 5-3) . Upon further 
examination of the pUFA-Hl clone by Southern hybridization analysis, 
it appeared that the cloned insert was not of plasnid origin. The 
radiolabeled clone did not hybridize to Southern transfers of plasmid 
DNA derived from strain H, or 16 other strains of X. c. pv. 
malvacearum . However, this 4.9 kb cloned insert of pUFA-Hl did 
hybridize to a 23 kb EcoRl fragment of the plasmid found in strain W. 
Additionally, the radiolabeled pUFA-Hl clone hybridized to a 4.9 kb 
EcoRI chromosomal fragment of 18 strains of X. c. pv. malvacearum , and 
also to another clone, pUFA-702 (33). pUFA-7C2 was independently 
derived and appeared to encode gene-for-gene avirulence on Acala Bj^. 
Thus, it was concluded that the pUFA-Hl clone was of chromosomal 
origin. 

Plasnid Origin of Replication and Mobilization 

Seven of the eight EcoRI fragments of the single 59.5 kb plasmid 
found in X. c. pv. malvacearum strain X were cloned into pUCD5 (Table 
5-4). All of the fragments were successfully mobilized into a 
spectinomycin resistant strain of X. c. pv. malvacearum strain N (N- 
SP^) in the presence of the helper plasmid, pRK2013, using MOPS 
minimal medium selection. In general, it appeared that the larger 
cloned fragments transferred at a slightly lower rate than the analler 
cloned fragments, however the transfer rate was not quantified. One 
of these clones, pLXF, carrying a 4.7 kb fragment, was found to encode 
plasmid transfer function (tra) , missing on the vector pUCD5 (14). 
That is, pXLF could be mated into X. campestris without aid of the 



62 



Table 5-4. Selection of plasmid replication genes in X. campestris pv. 
malvacearum strain X. ~ 





















Size 






Matings^ 










MOPS 


Kin 


POPS K^oSPinn 




Plaanid 


Insert 


no helper 


pRi:2013 


PRK2013 






pLXA 


17.0 kb 


few 




+ 








pLXB 


14.0 kb 


NT 




NT 


NT 






pLXC 


9.0 kb 


- 




+ 


few 






pLXD 


5.0 kb 


- 




++ 


++ 


(ori) 




pLXE 


4.8 kb 


- 




+ 


__ 






pLXF 


4.7 kb 


++ 




+++ 


_ 


(tra) 




pLXG 


4.5 kb 


- 




+++ 


_ 






pLXH 


0,5 kb 


— 




+++ 


few 







^ - is no growth, + = about 100 colonies, ++ = 100-300 colonies, +++ = 
swamped plate, NT = not tested. MOPS K30 is the MOPS minimal medium 
containing kanamycin (30 ug/ml) and POPS K30SP100 is the POPS 
canplete medium containing kanamycin (30 ug/ml) and spectincmycin 

(100 ug/ml). PRK2013 is a helper plasmid used in tri-parental 
matings (22) . pLXD contains the putative plasmid origin of 
replication ^ ( ori ) and pLXF contains a putative plasmid transfer 

(tra) function of the X. campestris pv. malvacearum strain X 
plasmid. ' 



helper plasmid. Since the vector pUCD5 is known to carry a 
mobilization function, it could not be concluded that the strain X 
plasmid is self-mobilizable, but it must carry at least some functions 
required for self-mobilization. 

PUCD5 is a broad host range vector with an origin of replication 
(ori) that functions poorly in X. c. pv. malvacearum . Matings on MOPS 
minimal medium provide a lengthy period of selection in which plasmid 
integration can take place, provided a homologous X. c. pv. 
malvacearum DNA fragment is cloned on the vector to allow 
recombination to occur. Rapid selection procedures using complete 
media containing antibiotics only rarely allows this integration to 
occur before the recipients die. vfnen transconjugants with X. c. pv. 
malvacearum strain X plasmid DNA were selected on POPS complete medium 
with antibiotics only the clone containing a 5.0 kb EcoRI fragment, 
pLXD, mated at a high frequency. All other clones either did not grow 
on POPS with antibiotics, or grew poorly. Because pLXD grew rapidly 
on selective medium, whereas other clones did not, it was proposed 
that the pXLD clone contained a functional origin of replication from 
the strain X plasmid. 

Transconjugants containing pXLD were further analyzed because 
their growth rate was highly variable. Eleven of the pLXD 
transconjugants were selected by colony size and grouped as large, 
medium or small. Plasmids from the transconjugants were extracted and 
analyzed for plasmid content. All, but possibly one transconjugant, 
were missing a 4.5 kb EcoRI fragment associated with the indigenous 
plasmid of X. c_. pv. malvacearum strain N-SpR (pig 5_2) . it was 



64 




23 



8> 



<Ii:k:::*i 



Figure 5-2. Transconjugants of X. campestris pv. malvacearum N-Sp^ 
containing pLXD (top) and hybrFdized to the plasmid probe pSa4 
(bottom) . The transconjugants were digested with the restriction 
endonuclease EcoRI (top). Lanes A and M, Lamda Hindi I; lanes B-E, 
large colony-size; lanes F-I, medium colony-size, and lanes J-L, small 
size transconjugants. All of the transconjugants, except in lane J, 
appear to have lost the 4.5 kb plasmid fragment seen'" in wild-type 
strains. Plasmids in lanes F-I and K-L appear to have had other 
plasmid rearrangements occur. All strains had an additional 8 kb EcoRI 
fragment present. The autoradiograph (bottom) shows the 
transconjugants probed with a pUCD5 related plasmid, pSa4 (93) . The 8 
kb fragment hybridized to the probe, but so did other plasmid 
fragments corresponding to those of X. campestris pv. malvacearum. 



65 



presumed that the homologous recanbination occurred between the pLXD 
clone and the indigenous plasmid. As a result, there was no EcoRl 
fragment which corresponded to the 13 kb pUCD5 vector, or the 5 kb 
pXLD insert. However, a new 8 kb EcoRI fragment was present. 

The plasmid DNA from the x. campestris transconjugants with pXLD 

was subsequently extracted and transformed into E. coli ED8767. The 

transformants were selected on Luria-Bertani medium containing 

kanamycin (25 ug/ml) , or ampicillin (30 ug/ml) . Resistance genes for 

these antibiotics border the EcoRI cloning site of the pUCD5 vector 

(14) . All of the transformed plasmids were 8 kb in size, which 

indicated that the new band seen in the x. c. pv. malvacearum 

transconjugants was, in fact, a new autonomously replicating plasmid 

(Fig. 5-3) . There was obvious rearrangement in the original pXLD 

clone which led to a 10 kb deletion. Southern hybridization analyses 

revealed the presence of DNA fragments from both pUCD5 and the insert 

of pLXD in this new vector, named pXD-1 (Figures 5-2 and 5-4). 

Discussion 
Of the twenty-five clonal isolates of X. c. pv. malvacearum 
derived fron 11 different nominal races, none of the clonal isolates 
corresponded to the expected disease reactions on the host 
differentials for that race. Likewise, there was no correlation 
between a given plasmid group and the observed pathogenic reactions on 
the 14 cotton differential lines. Although avirulence genes have been 
found on plasmids in other phytopathogenic bacteria (88) , no evidence 
has been found to link plasmids in X. c. pv. mialvacearum with race- 
specific interactions. In these studies, it appears the plasmids 



66 



H I 



M 



*ir^*p*" 



-,yi|.^§jl 







ii il-ii ii M M 



Figure 5-3. Transformation of E. coli EDS767 v;ith plasmid DNA from X. 
campestris pv. malvacearum transconjugants mated v/ith pLXD. Lane A, 
lamda H^dlll; lanes B-M, E. coli EDS767 transformed with DNA 
extracted from transconjugants mated with pLXD and selected on 
kanaraycin Luria-Bertani plates (see Fig. 5-2). The transformants were 
digested with restriction endonuclease EcoRI. Only tlie 8 kb EcoRI 
fragment was present; the smaller band was undigested DNA. 



67 



ABCDE FGH I A BC DE F GH 




Figure 5-4. Plasmid DNA of X. camcestris 



pv. malvacearum strain X 



digested with restriction endonucleases (left) and hvbridized to the 
plasmid probe pXD-1 (right). Lanes A and I, lamda Hindlll; lanes B-H, 
plasmid DNA from X. campestris pv. malvacearum strain X digested with 
BaiiHI (lane B) , Ball (lane C) , EcoRl (lane D) , Hindlll (lane E) , Kpnl 
(lane F) ,^ PstI (lane G) , and Sail (lane H) . The 5.0 kb EcoRI fragS^t 
seen m lane D is the fragment thoucrht to contain the origin of 
replication (ori). 



68 



served as useful markers for the genomic background. Plasmids in X. 
campestris generally appear to be stable and have in some instances 
been used as epidemiological markers, such as with the recent Florida 
outbreak of x. c. pv. citri (31) . 

However, five of the six strains with group I plasmids were more 
virulent than the other strains. All of the Upper Volta strains fell 
into this class, including scsne North American strains. The Upper 
Volta strains were derived from recently discovered races which were 
found to be virulent against all the known major genes for bacterial 
blight resistance (29) . The X. c. pv. malvacearum strain J, although 
an Upper Volta strain, appeared to be the exception and varied fron 
the other plasmids by the loss, or rearrangement, of a 4.5 kb EcoRI 
fragment found in all of the other 24 strains tested. Since the other 
group I strains are highly virulent strains, and known to be lacking a 
number of cultivar specific avirulence genes (33) , it may be possible 
that this 4.5 kb EcoRI fragment may serve sane positive function in 
pathogenicity. 

Nine avirulence genes have been reported cloned from X. c. pv. 
malvacearum , at least sane of which are chromosamally encoded (33) . 
With further testing, it is possible that plasmid-borne avirulence 
genes may be uncovered. In this study, approximately 50% of the 
plasmid sequences derived from an avirulent strain were tested, 
leaving it still possible to uncover avirulence genes in X. campestris 
pv. malvacearum . 

A high level of race-specific variability has been described for 
^' £.• pv. malvacearum (8) . Plasmids may be involved in genome 



69 

rearrangement as plasmids in P. syringae pv. phaseolicola have been 

found to integrate and excise from the chromosome causing 
rearrangement (79,92). m these studies, pUFA-Hl , carrying a 
chrcmoscmally encoded avirulence gene, was found to hybridize with a 
plasmid DNA from X. c. pv. malvacear-jm strain W. The evidence that 
plasmids in X. c. pv. malvacearum may be self-mobilizing points to 
their potential role in variability of the X. canpestris species. It 
is believed that the self-mobilizing, copper resistance plasmid 
identified in X, c. pv, vesicatoria facilitated the rapid build up of 
X. campestris strains resistant to field applied copper-containing 
conpounds (88). If avirulence genes are on self-mobilizing plasmids, 
rapid shifts in pathogenic races may occur, and the appearance of new 
races would be facilitated. 

The cloning of the X. c. pv. malvacearum plasmid origin of 
replication may allow for the future construction of other cloning 
vectors designed specifically for the study of genes in X. campestris . 
Sane of the currently used broad-host-range vectors appear unstable in 
X. campestris and are useful because cloned inserts integrate into the 
chromosome. Although preliminary experiments with the X. campestris 
ori resulted in vector rearrangement, possibly the resolved form may 
be applied to the construction of future vectors. 



CHAPTER SIX 

PATHOVARS OF XANTHOMONAS GAMPESTRIS ARE DISTINGUISHABLE 

BY RESTRICTION FRAGMENT LENGTH POLYMORPHISM* 



Introduction 
The species Xanthononas campestris (Pairmel) Dowson, a member of 
the family Pseudomonadaceae which is always found in association with 
living plants, comprises over 125 different pathovars (6). The 
pathovars have been named by "the plant from which first isolated" 
convention. Unfortunately, strains of some pathovars can be 
pathogenic on host species in different plant families (57); thus 
possible relationships between pathovars may go undetected, since 
there is no convenient way to determine pathogenicity for all possible 
host plants, nor is there a useful systematic method for narrowing the 
possibilities. Further, and despite a widely held view that all 
xanthononads are plant pathogens (85), non-pathogenic members of X. 
campestris are often found as ectoparasitic leaf colonizers 
(epiphytes) (26) and occasionally as endoparasites (65,71). Their 
inability to provoke a pathogenic plant response renders them 
unclassifiable in this system. The attention given to damaging plant 
pathogenic strains may not be taxonomically or ecologically warranted. 
Genes which function to provoke a pathogenic response may be entirely 
different from those which confer host selectivity and parasitic 
ability (32). Epiphytes and nonpathogenic endoparasites are known to 

* This chapter contains copyrighted material from International Journal 
of Systematic Bacteriologv. it is reprinted here with permission of 
the publisher. 

70 



i 



71 
be host selective (19,65). Host selectivity is thought to be a stable 
characteristic (84). since xanthomonads are always found in 
association with living plants, host selectivity should result in 
genetic isolation of a pathovar poEWlation. Over time such isolation 
and random genetic drift should produce distinctive and stable 
phenotypic characteristics and therefore, establish conserved genetic 
markers. 



„J Attempts to differentiate pathovars by methods other than 

j pathogenicity have included rRNA-DNA and DNA-DNA hybridization 

I (21,73), serology (1,94,95), phage-typing (37,59), and comparisons of 

I profiles resulting from plasraid and chromosomal DNA restriction enzyme 

,; digests (40,54), protein electrophoresis (70), and gas-chroraatography 

of fatty acids (65). Although all of these methods are useful for 

specific purposes, few have demonstrated utility to replace, clarify, 

or indicate further pathogenicity tests. 

We report here the use of restriction fragment length 
polymorphism (RFLP) analysis to differentiate pathovars of X. 
campestris. RFLP analyses have been widely used in the medical field 
to identify DNA fingerprints specific for inherited diseases or other 
genetic loci (2,11,61). These analyses have also been applied to 
plants (41), bacteria (38), and eukaryotic organelles (67,91). RFLP 
analyses allow the observance of genetic variation in organisms within 
defined regions of the genome due to DNA rearrangements or mutations 
which affect recognition sites for restriction endonucleases. 
Variation may be examined over a small (a few base pairs) or large 
(30-40 kb) stretch of DNA depending on the desired level of 



72 



polymorphiari detection. This variation can be observed for random or 
selected DNA target sequences within a species against a background of 
other genomic DNA fragments. In this study, RFLP analyses were 
applied to investigate the degree of genetic variation among 87 
strains of X. campestris, comprising 23 different pathovars. Seme of 
these results have appeared in abstracts (34,55), 

Materials and Methods 
Bacterial Strains 

The bacterial strains used in this study are shown in Table 6-1. 
These strains were isolated as plant pathogens, identified as x. 
campestris, and classified into pathovars according to the susceptible 
host plants involved. Strains were tested for pathogenicity after 
single colony purification. For some strains, no pathovar assignments 
were made due to the lack of known plant pathogenic responses. Broth 
cultures of bacteria were grown in a peptone-glycerol medium (per L: 
I 10.0 ml glycerol, 20.0 g peptone, and 1.5 g K2hT04; pH 7.2). The 
strains were camonly stored and maintained at -80 C in the same 
medium containing 15% glycerol. 
DNA Extraction 

DNA was extracted from bacterial cultures at mid- to 
late-logarithmic growth phase. Extraction of bacterial DNA for cosmid 
library construction was by a modification of the method of Silhavy et 
al. (87). A critical modification was to v/ash cells in 50 mM TRIS, 50 
rm NasEDTA, 150 mM NaCl, resuspend in the same buffer containing 150 
ug/ml proteinase K, add SDS (sodium dodecyl sulfate) to 1% (w/v) , and 



Table 6-1. Strains of X. campestris used for RFLP analysis. 



73 









J - " 




FatixTvar 


Strain 


Host 


LuLKtion 


Smrc^ 


JfentinoTDrBs caipescns patncvar 
alfalfae is 


M^icagD sativa 


Kansas 


D.L.StutEville 




ti. 


M. ^tJva 


Florida 


R,F;.,qtan 


arqemxes 


084-1052 


ArcBTcre iiExicarB 


Eloricfe 


DPI 


kegcni^ 


084-155 


teqcnia sp. 


Florida 


DPI 


caicestxis 


XI 


Brassica oleracEa (cafcfceqe) Cklahom 


this stix3y 




084-8Uy,U84-il36 B. deracm (cahtac^) 


FLoricfe 


■feds stu3y 




U«4-720 


B. oleracEa (br. sp:oats 


) Floricfe 


this study 




084-1318 


B. deracea (hroccoli) 


Eloricfe 


EPI 


carets 


13 


Caixus carota 


California 


R.E.Stall 


citri 


X59,}C70 


Citrus sp. 


Bcazil 


E.L.Civerolo 




XS2 


Citrus ^. 


J^Hn 


E.L.Civerolo 




X69 


Citrus sp. 


Ar^ntim 


E.L.dveroLo 




084-3401 


Citrus ^. 


Florid 


DPI 


cyaiDpsidis 


13D5 


Q'aicpsis tetracpTolcba 




CI.Kado 




X002,X005, 


C. tstragonoiabe 


ArizcTH 


J.Mihail 




XD16,XD17 


C, tEtragcHDlcte 


ArizoH 


J.Mihail 


dieffertedii^ 


084-729 


AntiJuriuB ^. 


Flori(fe 


IPI 




068-1163 


Dieffa±a±aa so. 


Floricfe 


EPI 


esaiLenti 


084-1093 


AcelTDsdxis esculentus 


Floric^ 


DPI 


qlyrires 


B-^3 


Gycix^BwsK 


Brazil 


W.F.Efett 




1717 


G. ITHX 


Africa 


W.F.Efett 




17915 


G. ITBX 


- 


W.F.Ffett 




S-9-8 


G. ITBX 


Wisocrsin 


W.F.Efett 


hefer^ 


084-1789 


fefera helix 


Eloricfe 


EPI 


holcicola 


}(n66 


Sdrxriin vuloare 


Kansas 


T„n aflin 


macalifoliigartfeii 


3e 084-6166 


(Srosnia sp. 


FIari(3 


DPI 


iraLvac^axm 


D,M,N,0,U,V, 


Gossv'piun hirsuton 


Ifexas 


i±iis sta3v 




X,y,Z,TX84 


G. hirsoton 


Ifexas 


this strriy 
M.Es^nberg 




A,B,E,F,G,H 


G. hiraitun 


Ck].ahanH 




Chl,a)2 


G. hirsuton 


0-^3 


L.S.Bird 




Hv^ 


G. hirsutun 


Uffer Volta 


L.S.BiE3 




Su2,Su3 


G. hirsuton 


arbn 


L.S.Bird 




EL79 


G. hirsuton 


Florida 


this strriy 


iTBngifer^iaaic^ 


084-U6 


tergifera inaica 


FLoricfe 




DPI 


nigraraculans 
pelargonii 


084-1584 
084-190,084-1370 


Arctiun lacpa 
Cferanior, ^, 


FLoricfe 
Eloricfe 


DPI 
DPI 


TTBseoii 


EKll,:^ii25,}$jEll 


Eteseolus vulqaris 


N&ra^ca 


M-S±iustEr 




:^,xpn 


P. vai.q3ris 


Mscxxsin 


A.W.^ttler 




82-1,82-2 


P. vuiqaris 


Florida 


R.E.Stall 




LB-2,S3-3B 


P. vulgaris 


iM^iraska 


A.K.Vid3ver 




XP2 


P. vulgaris 


Ifev York 


J.A.Iauraxe 




XEKIL 


P. vulgaris 


Kansas 


J.L.IfiQ±l 




XPnlF 


P. vul^ris 


MissGuri 


tMs study 




XE^-CPI 


P. vulqaris 


- 


thiis study 



74 



Table 6-1 (continued) . 



E&tiKwar 



pisi 

pDlnset±iicaLa 
pcuni 

transluce:^ 
■v^sicatoria 



vlgnicola 



vitians 
zinnis 



unknonn 



X. fragari^ 



Strain 



XPl 

083-6248 

084-1793 

X1I05 

B-3 

75-3 

A81-331,C-1, 

CK-1, 

»^19,aG, 

432,82-38 

ICEB164 

084-1944 

084-1373 

084-3928 

084-4348 

08:^2057 

084-2848 

084-1590 

251G,084-480 

084-6006 

XEcal 



Host 



Piam sativum 
Hpxdpia pulcrerriiTB 
Pnjrus ^. 
Hcccfeun ^. 
Cteigjn annxtn 
Lycc^ETsicon esaJLentuTi 
Vigna unguiculata 
V. ungiirulata 
V. urgaiculata 
V. urgoicaLata 
latixa ^. 
Zinnia 



rhilofedrcn ^. (dief fsi. ) 
E^tsia ^. (1-Bcfer^) 
Alocasia ^. (vitians) 
Syrgcniim ^. (vitians) 
Cissus ^. 
Ebcrpus ^. 
StpatieTs ^. 
Jasninium ^. 

Fi:agaria ^. 



Locaticn Scarce^ 



J^Bn 


MODto 


Flacida 


EPI 


Florida 


DPI 


^tnhat■H 


D.Sfends 


Florida 


R.E.Stall 


Eloricfe 


R.E.Stall 


feorgia 


R.D.Gitaitis 


feccgia 


R.D.Gitaitis 


Cfeorgia 


R.D.GitBitis 


Cfeargia 


R.D.GitBitis 


- 


R-E.Stall 


Eloricfe 


DPI 


Eloricfe 


DPI 


Florida 


DPI 


Eloricfe 


DPI 


ELoricfe 


DPI 


Florida 


DPI 


KLoric^ 


EPI 


Elari(^ 


DPI 


Eloricfe 


DPI 


Florida 


R.E.Sti=i11 



^ DPI. T Florida Department of Agricultural and Consumer Services 
Division of Plant Industry, Gainesville, Florida. services. 



75 

heat for 1 hr at 50 C. The sample was then extracted 2 times with 
phenol/chloroform/isoamyl alcohol (25:24:1; phenol equilibrated to pH 
7.8 with 0.1 M TRIS), and the DNA spooled out from the aqueous phase 
after adding sodium acetate at 30 mM and 2 volumes 95% ethanol. This 
DNA was then washed in 70% ethanol. Extracted DNA was resuspended in 
TE (10 mM Tris(hydroxYmethYl)aminomethane (TOIS) , 1 mM disodium 
ethylenediaminetetraacetate (Na2EDTA) containing 20 ug/ml DNase-free 
pancreatic RNase; pH 7.6). 
Agarose Gel Electrophoresis 

Approximately 5 ug of DNA was digested with EcoRI or BanHI as 
specified by the manufacturer (Bethesda Research Laboratories). 
Digested DNA samples were run in a 20x25 cm agarose gel (0.6%; type 
II, Signa) in TRIS-acetate buffer (40 mM TOI^acetate, 1 mM Na2EDTA; 
pH 7.6) with electrophoresis of gels set at 35 V for 14-15 hrs. 
Fragtients were visualized by ultraviolet irradiation (302 ran) after 
staining agarose gels in ethidium bromide (0.5 ug/ml). Photographs 
were taken using Polaroid Type 55 (or Type 57) film and a yellow 
filter (Tiffen no. 12) . After the gels were photographed, the DNA was 
transferred to nitrocellulose by the method of Southern as described 
by Maniatis et al. (66). Restriction fragment sizes were estimated 
using DNA molecular size standards of lambda phage DNA digested with 
Hindi II. 

DNA Probes 

The DNA probes used in this study were derived from a genomic 
library of X. c. pv. citri strain 3401 constructed into the modified 
coanid cloning vector pUCDSB, which was the vector pUCD5 (14) with a 2 



76 



kb B^nHI fragnent deleted from it. DNA of strain 3401 was partially 
digested with the restriction enzyme J^i and fractionated by size on 
a 10-40% (w/v) sucrose step gradient (66). The pUCDSB vector was 
digested with the restriction enzyme BamHI and treated with calf 
intestinal alkaline phosphatase (Boehringer-Mannheim) . Vector DNA and 
large fragnent size fractions (25-40 kb) from strain 3401 were mixed 
and treated with T4 DNA ligase (Bethesda Research Laboratories) and 
recombinant coanids were packaged in vitro with extracts prepared and 
utilized according to Scalenghe and Hohn's protocol II as described by 
Maniatis et al. (66). Transduction with Escherichia coli HBlOl was 
performed using top agar on Luria-Bertani medium containing kanamycin 
(35 ug/ml). Cloned DNA fragments of strain 3401 in the vector 
averaged 27-38 kb. 
DNA Hybridization 

DNA clones used as probes were radiolabeled using 30-60 uCi of 
32p_deoxycytidine (Du Pont NEN) with use of a nick translation kit 
(Bethesda Research Laboratories) and separated from low-molecular 
weight nucleotides on a mini-column of Sephadex G50-100. Southern 
blots were pre-hybridized in plastic bags for 4 hr and hybridized with 
the DNA probe for 15-18 hrs at 68 C. The pre-hybridization and 
hybridization solutions were those described in Maniatis et al. (66). 
Following hybridization, membranes were washed once in 2X SSC, 0.5% 
SDS (IX SSC is 0.15 M NaCl, 0.15 M sodium citrate, pH 7.0) and washed 
once m 2X SSC, 0.1% SDS at ambient temperature, and washed two times 
in O.IX SSC, 0.5% SDS at 68 C as described by Maniatis et al. (66) for 
'stringent' conditions. Nitrocellulose membranes were then air-dried 



77 

and exposed to Kodak X-Qnat AR film at -80 C in cassettes with 
intensifying screens. Hybridization of the probes to individual 
strains of X. campestris was repeated at least three times. 
Restric tion Fragment Patterns and Densitometry 

From developed autoradiographs, DNA fragment sizes and profiles 
were determined and hybridization signals measured using a Gilford 
Response spectrophotometer equipped with an autoradiograph scanner. 
Routine scans were done at 600 nm wavelength with 0.5 irni aperture and 
0.5 nm bandwidth setting. Scanning data was stored by computer. 
Using above information, canparison were made among strains of a given 
pathovar and also strains derived from different pathovars of x. 
campestris . 

Results 
Agarose Gel Electrophoresis 

For a given pathovar, the banding patterns appeared very similar 
and poliTOorphian in these strains was apparently limited (Fig. 6-1). 
In all, DNA from 87 strains of X. campestris , representing 23 
different pathovars and including some with no known pathovar status 
were digested with restriction enzymes EcoRI and BamHI and separated 
by size on agarose gels. Based on restriction fragment patterns 
alone, it was possible to visualize DNA variability among the strains, 
but recognition of particular DNA fragment banding patterns was 
difficult due to the numerous DNA fragments involved (Fig. 6-2) . Some 
of the apparent variation was due to the presence of plasmid bands, 
which appear brighter against the chronoscmal bands due to their 
relatively higher copy number within each cell. 



78 



ABCDE FGHI JK LMNOPQRST 




Figure 6-1. Genomic DNA of strains of X. campestris digested with the 
restriction enaonuclease EcoRl. Lane A, probe XCTl; lanes B-K, X c 
P^- ^^seoli; lane B, JL; lane C, LB-2; lane D, Xph25; lane E, EKirJ 
lane F, 82-1; lane G, 82-2; lane H, X?a; lane I, Jp2; lane J, XP2-l' 
lane K, JF; lanes L-M, X. c. pv. phaseoli var. fuscans; lane L, SC-3B 
lone M, Xp.ll; lanes N-0, X. c. pv. alfalfaeTTiH^ N, KS; lane 0, 
nRr?o?p Z' -%-T ^7' £^2Eestris; lane P, XCI (cabbage); lane Q, 
084-1318 (broccoli); lane R, 084-720 (brussels sprouts); lane S 
unknown X. c. G65; lane T, probe XCTll. 



79 



ABCDE FGH 



JI^LM N O P o p c; T 




Figure 6-2. Genomic DNA of strain >; ^-f v ^ 

restriction endonuclease Ecori Lanp a' ^^^^estris digested with the 
alfalfae FL; lane C X 7^ n/ ^ ' ^"""^^ ^^^^' ^^"^ B, X. c. pv. 
i^ilHs 084-809; li; 1 ? .^^fn^ae 084-155; lane D, xT c" pv 
cyarnopsidis X002- lane r v - - P^ • £^££I^ 13; lane F, X. c. pv 

i^^l^^^^^-Ii^ S-9-3''line-i 'x' l^^^^^f^^^ 0^8-11637 lane^^ 
c. PV. -^nMTigard^"^^^ P- ^^^£ol^ Xh56; lane J, x. 

^79; laH^LTxr^r^^T^inrgonii 08^190 lane-A 'x' ^"^' ^H^^^^^eanjn 
var. fuscans SC-3B- lan e m v — ;; '• "' -■ ^- P^- Phaseoli 

transWHTxilOS; lane P x "c^ "^v ""* ^ ''''^' '""^ °' X-^^T^^ 
^^^n^iinl^ola SN2; lane R, ~X ~c "S^' iS^^^^^ '^'' 1^"^~Q^~X. c. 
pv. HnHIiFo-84-1944; i^^e 't," prl'oe'^XC^rf^^ ''''^'''' '""^ "' ^- ^- 



80 



MA Probes 



Two cosmid DNA clones, XCTI and XCTII were randomly selected from 
a genomic library of strain 3401 for use as DNA probes. xCTl carried 
a 30 kb insert, and XCTll carried a 37 kb insert, as determined by 
adding up the cloned DNA inserts from restriction endonuclease 
digests. There were no detectable plasmids in strain 3401, thus the 
coanid clones are assumed to contain chromosomal fragments. Assuming 
a 3333 kb genome for X. campestris (47) , each cosmid probe represents 
approximately 1% of the total genome. 
DNA Hybridization 

Autoradiograp^s of Southern blots hybridized against either of 
the two DNA probes revealed conserved DNA fragments within each 
pathovar. For example, in figure 6-3 (lanes B - K) are shown the RFLP 
pattern from 10 different strains of X. c. pv. phaseoli from different 
geographic locations. At least five EcoRI cut DI^A fragments which 
hybridized to xCTl appeared to be conserved in this pathovar. m 
lanes L and M are the RFLP patterns of two strains of X. c. pv. 
^^as^li var. fuscans. These X. c. pv. phaseoli var. fuscans strains 
also attack beans (and hence have the pv. phaseoli designation), but 
are biochemically distinct (6). They produce a dark, melanin-like 
pignent in nutrient media and in these tests are clearly genetically 
distinguishable from the other X. c. pv. phaseoli strains. Conserved 
DNA fragments were also seen for X. c. pv. alfalfae (Fig. 6-3, lanes N 
and 0) in which six of the up to 15 fragments appeared identical. 
Likewise, in four different strains of x. c. pv. campestris (Fig. 6-3, 
lanes P - Q, Fig. 6-5, lane D) isolated from three different crucifer 



81 



ABCDEFGHI JKLMNOPQRST 



12.2 
11.2 



7.9 

5.6 
5.1 



3.2 



2.2 



1.6 




Figure 6-3. Genomic DNA of strains of X. campestris digested with the 

2- ?anP n v^.- 1 '^'^' ~ -' P^' ^^^^^^'^ lane B, JL; lane C, L&- 
L l^ne T "^^pV r^ ^' ^^^'' '^"^ ^^' ^2-1; lane G, 82-2; lane H, 
Xpa, lane I, XF2; lane J, XP2-1; lane K, JF; lanes L-M, x c pv 
|-|^ var. fuscans; lane L, SC-3B; lane M, Xpfll; lanes torx.^l'. 

lane P XCI (cabbage); lane Q, 084-1318 (broccoIi)7 lane ' r, 084-72 
(brussels sprouts); lane S, unknown X. c. G65; lane T, probe XCTll. 



sources (cabbage, broccoli, and brussels sprouts) at least four of 
over 10 frag:nents appeared to be identical. sizes of hybridizing DNA 
fragments are given in Table 6-2. in figure 6-4 is shown the RFLP 
patterns revealed using clone XCTII. m each case, a different RFLP 
pattern is shown, but the same genetic relationship is revealed. 
Other sets of conserved DNA fragments within a pathovar were seen 
among seven strains of X. c. pv. vignicola (identical pattern) , seven 
strains of x. c. pv. c yamopsidis (nearly identical), and at least 
three strains each for x. campestris pvs. cltri, glycines , 
inalvacearum, pelargonii , and vesicatoria (not shown) . 

DNA of strains representing other pathovars of x. campestris were 
also digested with either EcoRi or B^i, and hybridized with either 
XCTl or XCTll. The pathovars of x. campestris included in these 
experiments were x. campestris pvs. alfalfae , begoniae, cam^stris , 
2a£otae, citri, cyamopsidis, dieffenbachiae . esculenti, glycines , 
hederae, holcicola , macul i fol i i garden ... . malvacearum, 
mangiferaeindicae , nigromaculans . pelargonii , phaseoli, oisi, 
translucens , vesicatoria, vignicola, vitians , and zinniae. Also 
included were sane strains isolated as pathogens of an Alocasia sp. , 
Cissus sp., impatiens sp., Fatsia sp., Jasrr.ine sp., Argemone sp., and 
a Euonynus species for which no pathovar designations were available. 

The results of hybridization of probe XCTl to DNA of different 
pathovars of x. campestris digested with EcoRI are shown in Figure 6- 
5. The clone XCTl appeared to contain DNA fragments of strain 3401 
which were able to demonstrate variability of restriction fragnent 
patterns among different pathovars of x. campestris . Variability in 



82 



p^^^'i^-r^s-^yS!-^^^^^^^^ 



83 



^ ^^?, ^^2, 7.9, 5.6, 5.1, 3.2, 2.2, 



M 



P 
Q 
R 

S 

T 



R 777 7o o n ' •''' ^•^' ^-2' 2.2, 1.6, 1.0 

C U± lit lit l-t' f-:' f-5- '■?' ^•°- 3-A 3.3. 3.1 

D 

E 13 

F 



14.4, 13.3, 10.4, 7.6, 6.4 4.6,' 4:3; 4 3 7 ''3 .7 
14.1, 13.1, 10.4, 7.5, 6.4, 5.2 4 6 43 4"o Vl' l'^ 

13.7, 13.1, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1, 
13.7, 13.1, 11.6, 10.4, 8.9, 5.2, 5.0, 4.6, 4.3, 4.0, 3.7, 3.1, 



I B B it it =t i- "■ -; -; 3t \:i 



L 11.9, 8.4, 7.1, 6.7, 4.5 



17.0, 15.3, 10.0, 8.4, 4.5 

^ £'3?- "- "- "- -^; ^:?; 't i-k It ?t k?. 



*!' ^«"n' ^.^'^ ^^•^' ^2-^' ^2.1, 8.0, 4.6, 
4!'3!S ' ^^•^' '^-^^ ^^•^' 5-^' 7-3' 

f.t'h'Xf''' ''■'' ^'''' '-'' '■'' '-'' '-'^ ^-2. 4.4, 3.9, 
2'4!'2.'2?1.9'°-'' '-'' '■'' '■'' '-'' ^-3, 4.6, 4.4, 3.9, 3.3, 
18.6,^15.8, 12.7, 10.8, 10.2, 9.1, 8.4, 7.7, 6.7, 5.6, ,.,, 3.3, 

11.3, 8.0, 6.9, 3.4, 2.5, 1.6 

17.2, 15.3, 14.3, 10.5, 8.6, 7.2, 6.8, 5.8 



84 



ABC p E F G H I J K LMNOPQRST 





■■<Bls 



n^,. S*#v 



1^ 



.:/^-^:-.. :li|j5;. 




XDa- lanp T vdo. i ^ ' ^' ^^"^' ^^"^ "2, 82-2; lane H, 

^dne F, xci (cabbage); lane Q, 084-1318 (broccoli)" lane R 'nR4 79 n 
(brussels sprouts); lane S, unknov^ x. c. G65; lane T, prole xcSL 



85 



ABCDE FGH 



12.2 
11 J 



7.9 

5.6 
5.1 



3J2 



2:2 



1.6 




J KLMNOPQRST 




rlltrlctlon SZ^l^L%^r'7 °'-^^ ^^^^-tris digested with the 

A, probe X.r^'^TTiX^'^^S^i^'^^r'' ^^"^ 
begoni ae 084-155; lane n ~y 7^ t^-, — " . ^^' -^^® ^' >^- c. pv. 

dieffiHb^iae 058^1163 • larl" PV - S^nopsidis X002; lane G, X. c7 pv. 

c. pv. holcI Fola Xh66.'lane J 7 I" ''''" ^^^^^^^ ^-'-'' ^^"^""^' ^■ 
6166- l ane K v V ^ . ' - -* P^* maculifoli j garden iae Qsil 

pv. pis XPl!- I'ani 0- -"^ c ^^^ "'T' ^^^^^^ ^^-^^' lan^l^TxTH: 

vesiitoria 75-3 lane ^5' l' "^^ ^ranslucens X1105; lane P, x. c? pv. 

^ tians IC PB164>' W S F F T" -^-^"^.^"'-^ ^^^r lane R, x. c. pv. 

___^ «ib4, lane S, X. c. pv. zinniae 084-1944; lane "t, "probe 



86 

R^P pattern appeared as ^.alitative and quantitative differences 
-ong hybridization profiles of various sized DNA fra^ents in the 
strains. For instance, the weak hybridization signal seen from X. c. 
pv. translucens (fig. 6-5, lane 0) suggests that this strain is 
distinct from the other pathovars conpared, at least over the 1% of 
the genome contained on the clone XCTI. m other pathovars such as X. 

c^- pv. alfalfae (fiq. 6~^ lanf^ n\ v ^ -l 

b. t. D, lane B) , x. c. pv. begoniae (lane C) , x. c. 

pv. c^^anopsidis (lane F) , x. c. pv. malvacearum (lane K) , X. J pv. 
mm^ (lane Q) , and X. c. pv. vitians (lane R) where sJe DNA 
fragments had strong hybridization signals, a close relatedness with 
portions Of DNA in the co^id clone is suggested. Seme of the DNA 
fragments from different pathovars which hybridized most strongly to 
the probe were also of identical size with the EcoRi fragments of the 
XCTI clone. As expected, DNA of X. c. pv. citri strain 3401 digested 
with ECORI, or Bam«I (34,55) contained fragments which corresponded to 
DNA fragments of the XCTI probe (Fig. 6-5, lane A). Sizes of 
hybridizing DNA fragments are given in Table 6-3. 

From observations using different enzymes and different probes, 
the same patterns emerged, indicating that the basic chromosome 
structure of x. c^npestris can be used to differentiate strains into 
specific RFLP types. Although a pathovar may contain more than one 
type of variant (e.g. x. c. pv. ^^aseoli ) , all strains of a given type 
appear to exhibit the same host selectivity. The RFLP pattern is 
highly distinctive for each type, and can be used to unambiguously 
assign an unknown sample to a pathovar, using simple visual 
comparisons of pattern against a set of known samples, such as those 
given in Figure 6-5. 



87 



Table 6-3. Sizes of DNA fragments from Xanthononas campestris genomic 
digests (EcoRI) which hybridized to the XCTl asIA probe". See Figure 6- 
5« 



A 12.2, 11.2, 7.9, 5.6, 5.1, 3.2, 2.2, 1.6, 1.0 

B 20.5, 18.8, 15.3, 14.2, 13.5, 12.0, 7.3, 6.2, 5.1, 4.6, 3.7, 3.4, 

3.1 
C 18.7, 12.1, 9.5, 8.6, 8.2, 7.7, 6.9, 6.7, 6.2, 6.0, 4.8, 4.2, 

4.0, 3.4, 3.0, 2.9, 2.5, 2.3 
D 19.4, 16.4, 14.4, 13.1, 11.1, 10.5, 9.2, 8.5, 7.7, 6.7, 5.6, 4.3, 

4.0, 2.2, 1.9 
E 17.9, 12.9, 4.1, 3.5, 3.0, 2.3, 1.5 

F 16.2, 11.5, 9.5, 8.7, 5.7, 4.0, 3.6, 3.1, 3.0, 1.4 

G 14,2, 11.1, 4.5, 3.9, 3.5, 3.0 

H 12.2, 10.2, 8.3, 7.9, 6.5, 5.5, 4.2, 3.1 

I 12.9, 12.0, 7.6, 7.1, 6.5, 4.7, 3.0, 1.7 
J 15.6, 14.2, 9.4, 7.9, 6.3, 5.0, 4.6, 4.0, 3.1, 2.5, 2.2 

K 20.9, 17.1, 16.2, 15.5, 14.2, 11.0, 10.4, 9.2, 8.2, 7.5, 6.7, 

4.2, 4.0, 3.3, 3.1, 3.0 

L 19.4, 12.7, 10.4, 9.0, 7.9, 7.7, 6.8, 5.7, 4.2, 3.9, 3.0, 2.7, 

2.3, 1.9, 1.5, 1.4 

M 12.0, 8.4, 7.0, 6.8, 4.5 

N 15.9, 8.4, 4.9, 3.4, 2.1, 1.9 

9.2, 4.4 

P 13.3, 9.2, 7.2, 6.9, 6.2, 4.6, 3.8, 3.1, 2.1 

Q 11.0, 8.5, 7.0, 6.3, 4.2, 3.2 

R 16.7, 14.2, 12.9, 9.9, 7.4, 6.9, 6.3, 4.8, 3.8, 3.5, 3.2, 2.6, 

2.4, 2.2, 1.6 

S 12.9, 12.5, 9.5, 7.7, 6.5, A. A, 4.0, 3.8 

T 17.6, 15.6, 14.7, 10.5, 8.6, 7.2, 1.4 



88 



Discussion 
The only widely accepted and most practical method for 
differentiating pathovars of X. campestris is to inoculate a plant 
suspected as the host for that pathovar. This practice can sonetimes 
be tedious, time consuming, subjective and subject to a surprising 
number of artifactual influences. It is not known whether host 
selectivity is unstable as suggested by Dye (23); stable as suggested 
by Schnathorst (84) or even taxoncmically significant. Although the 
classification is thought to be useful, it can be highly misleading 
since emphasis is placed upon one characteristic - pathogenicity. If 
only one or few gene differences were involved in host selection, then 
differentiation by pathovar could be highly misleading, at least in 
the sense that a given group of strains might be capable of attacking 
more than one host in some cases. Alternatively, two or more strains 
of relatively unrelated groups could be cataloged together because 
they happen to attack the same host. This latter situation appears to 
be the case with X. c. pv. phaseoli and X. c. pv. phaseoli var. 
fuscans (see Fig. 6-3) , and with Florida strains of X. c. pv. citri 
(34). 

To address these problems, v.'e evaluated the potential of using 
DNA sequence variation of X. campestris for strain classification 
purposes. DNA sequence variation was first examined by digesting DNA 
with restriction enzymes and visualizing directly the resulting 
fragments on ethidium bromide stained gels. These stained gels were 
useful for side-by-side comparisons of restriction fragments for 
samples run on the same gel, but comparisons between different gels 



89 



were more difficult. Subsequently, visualizing and comparing 
variation in bacterial genomes with cosnid clones carrying 30-38 kb 
cloned X. campestris genome fragments was simplified. In both cases 
variation was revealed by alterations in the sizes of visualized DNA 
restriction fragments (RFLPs) . The use of DNA probes derived from 
chromoscmal DNA fragments also alleviated seme of the difficulty in 
comparing stained gels with the sonetimes present polymorphism of the 
higher copy number plasmid DNA fragments which can occur in the 
background . 

Seme of the selected clones tested as DNA probes appeared to be 
useful for identifying DNA sequences conserved within a given 
pathovar, while others appeared to identify DNA sequences which were 
highly conserved at the species level. The DNA which was considered 
conserved at the species level was represented as banding patterns 
V7hich were nearly identical among all strains over the several 
pathovars tested. Examples of DNA sequences which are known to be 
highly conserved are rRNA encoding genes (38,83). If smaller DNA 
probes containing known genetic loci such as those associated with 
genes for pectate lyase and protease (18), or avirulence activity 
(33,88) were tested, these smaller, more defined DNA probes could be 
used to determine if such genetic loci were highly conserved or 
variable. With larger, randomly selected probes, the likelihood was 
increased for detecting strain- or pathovar-specific variation. The 
large (greater than 30 kb) size of the probes used in this study 
allowed detection of both highly conserved and variable regions. 



90 



Using RFLP genomic blots, it appeared that phylogenetic 
relationships between the full spectrum of described pathovars of X. 
campestris might be determinable. Some mathematical approaches toward 
determining phylogeny based on restriction cleavage sites has been 
proposed (58,74). DNA probes have been used to make phylogenetic 
comparisons among the relatively conserved mitochondrial (67) and 
chloroplast (91) gencsnes. The variation seen among the different 
pathovars of X. campestris appeared within the range that is usefully 
distinguishable with this test. Only limited amounts of variation can 
be usefully distinguished, such as that occurring within a species. 
Significantly, the RFLP groupings closely corresponded with the 
pathovar groupings, strengthening the taxonomic significance of this 
classification. All strains tested were readily grouped by RFLP 
phenotypes, and the classification based on RFLP patterns correlated 
very well with the classification based on pathogenicity. This 
technique may provide a more convenient means of classifying these 
bacteria. in addition, unexpected taxonomic relationships between 
pathovars might be revealed. The potential pathogenic range of each 
RFLP group v;ould be of obvious value. 

By comparing observed RFLPs among strains of X. campestris using 
selected DNA probes, it was possible to identify unknovm strains when 
known standards were included. Additionally, strains previously 
undescribed could be classified as being related to knovm pathovars. 
For instance, similarities were found between undescribed strains 
isolated from a Alocasia sp. and Argemone sp.; and a Cissus sp. and 
Jasmine species. The strains isolated from a Philodendron sp. and 



91 



Dieffenbachia sp. both corresponded to that of X. c. pv. 
dieffenbachiae . The other unknown strains tested appeared different 
from one another, and did not appear similar to any of the other X. 
campestris pathovars as characterized by the DNA probes. Finally, 
this technique may provide a basis for the classification of 
nonpathogenic and epiphytic xanthomonads. Experiments can now be 
conducted to determine whether epiphytes are all basically similar or 
form diverse groups, whether they are restricted in host range or if 
they are nonpathogenic, but closely related to described pathovars. 
Although this approach provides another classification scheme for 
strains, it appears to be compatible with the currently recognized and 
useful pathovar naming scheme, which relies primarily on plant- 
specificity. Although certain pathovars may need to be redefined, 
this work supports and helps validate the natural taxoncmic groupings 
provided by the pathovar naming system. 



CHAPTER SEVEN 
SUMMARY 



Intraspecific Variation 
Analyses of intraspecific variation were conducted using conmon 
microbiological techniques, including SDS-PAGE of total proteins. 
These methods revealed strain-specific, but not pathovar-specif ic, 
variation. Attempts to differentiate strains into pathovars of X. 
campestris were not successful by these means. The j±iysical primary 
structure of plasmids and chromoscmes, as revealed by restriction 
fragment length polymorphian (RFLP) , was useful in revealing pathovar- 
specif ic variation. Strains of unknown plant origin can be at least 
tentatively classified by pathovar. Additionally, non-pathogenic 
epiphytes may be classifiable (epivars?) using these methods. 

Plasmids 
Plasmids appear useful to differentiate among some of the 
pathovars of X. campestris . In comparing strains of X. c. pv. 
malvacearum with race-specific interactions, no correlation was 
observed between plasmids grouped by RFLP patterns and cotton host 
differentials. Furthermore, plasmid DNA fragments were cloned in 
broad-host-range vectors, mobilized into virulent strains and tested 
for race-specific avirulence. Although not all of the possible 
plasmid sequences were tested, it is still possible that avirulence 
genes may reside on plasmids of X. c. pv. malvacearum. However, sane 

92 



93 



host-range specifying functions may be plasmid encoc3ed. Strains of X. 
c. pv. phaseoli and X. c. pv phaseoli var. fuscans have the same host 
range, but differ significantly in colony appearance, physiological 
tests, and chromoscmal RFLP analyses. However, plasmids from these 
strains appeared highly related by restriction fragments profiles and 
Southern hybridization. Thus, these strains are quite distinct 
biochemically and chromosomal ly, but similar as characterized by the 
plasnids and by host range. This correlation between host range and 
plasmids allowed the use of plasmids and cloned plasmid fragments as 
epidemiological markers for detection and rapid identification of X. 
c. pv. citri strains in Florida. Plasmid RFLP analyses are limited to 
those strains which carry plasmids; only 95 of 151 strains of X. 
campestris tested, or 11 of 26 pathovars tested had plasmids in all 
strains. 

Restriction Fragment Length Polymorphism 
Chromoscmal DNA RFLP analyses proved to be a reliable method to 
differentiate X. campestris by pathovar, at least for those pathovars 
tested. This method provided a means to resolve taxonanic 
ambiguities. For example, X. c. pv. phaseoli and X. c. pv. phaseoli 
var. fuscans , appear quite different by chromoscsnal RFLP analyses, and 
share a common host. By contrast, nearly all 100 strains tested from 
24 pathovars had RFLP patterns characteristic for the pathovar, and 
very little intrapathovar variation was present. 

The use of RFLP methods has already been useful in other studies 
as an applied tool in uncovering the origin of strains of the citrus 
canker pathogen found in Florida (Gabriel et al., unpubl.) . An RFLP 



94 



survey of the knovm type strains of X. c. pv. cltri and those isolated 

in Florida allowed for the separation of these bacteria into four 

identifiable RFLP groups. Strains belonging to the "A group" isolated 

frcra Brazil and Japan were related by RFLP analysis. 1*he strains 

isolated in 1986 from citrus in backyards of the Tampa-Bradenton area 

were also of the "A group". Strains belonging to the "B group" and "C 

group" (isolated from Argentina and Brazil, respectively) appeared to 

be related by RFLP analysis, yet could be differentiated. A Mexican 

cancrosis strain was indistinguishable from the "B group" baseS on 

RFLP analysis. Other strains of X. c. pv. citri isolated in Florida 

did not appear to be related to those strains placed into the "A 

group" or "B/C group". The Florida nursery strains appeared to 

constitute yet a third RFLP grouping. This group was highly 

polymorphic, and some strains appeared to be related to other 

pathovars of X. campestris , strengthening the "mutation/endemic" 

theory for some occurrences of citrus canker (Gabriel et al., 

unpubl . ) . 

RFLP methods are powerful taxonomic tools to distinguish major 
differences in strains assigned to the same pathovar simply because 
they share a common host. For instance, the differentiation of 
strains of X. campestris pathogenic to crucifers, such as X. c. pv. 
campestris and X. c. pv. armoraciae , is at times difficult. X. c. pv. 
campestris is generally a systemic pathogen, whereas X. c. pv. 
armoraciae is usually localized, but several strains showing 
intermediate symptoms have been identified (99). In preliminary 
experiments the RFLP methods were able to differentiate between the 



95 



strains. Surprisingly, differences could also be detected among the 
strains assumed to be X. c. pv. campestris , where seme of the more 
virulent strains belonged to a separate RFLP grouping. 

Because RFLP groupings of X. campestris correlated well with host 
range, it seems possible that this association could also be valid 
with other bacterial systems. There is a great interest in the use of 
microorganians as biocontrol agents. In many cases, attempts to 
exploit microorganisms which exhibit strong in vitro antagonism have 
failed due to the poor survivability of the agent in plants. For 
instance, P. fluorescens has been used in developing rhizosphere 
conpetent biocontrol agents against fungi (44,63,78,100). Strains of 
P. fluorescens , also a member of the same bacterial family as X. 
campestris , is poorly differentiated; and as part of the soil 
rhizosphere is also poorly defined. RFLP studies could be used to 
identify members of P. fluorescens which have the desired host range, 
thus, facilitating the development of more effective biocontrol 
agents. 

Conclusions 
The extent to which we devise taxonomic distinctions to 
differentiate among organisms remains a function of their utility to 
humankind. Their relavance to "natural groupings" is in the minds of 
pathologists or microbiologists v;hose perception may be somewhat 
different fran their Creator. The variation which is perceived as 
significant by pathologists to differentiate strains of X. campestris 
is host-specific pathogenicity. The variation which is perceived as 
significant by geneticists may be chranosanal RFLP patterns. The 



96 
taxoncmic validity of grouping strains by pathovar was strengthened, by 

the observation that, in general, strains v/ithin a pathovar fall into 

a single RFLP group. Although plant tests are still necessary as a 

confirmation of host pathogenicity, mutants which are no longer 

pathogenic and non-pathogenic epiphytes may now be classified in one 

classification system. The analyses conducted using common 

microbiological techniques, including SDS-PAGE of total proteins, 

revealed the relatedness of strains at the species level. Attempts to 

differentiate strains into pathovars of X. campestris were not as 

successful by these means, as variation was unresolvable at this 

level. with the incorporation of additional plasmid and RFLP 

analyses, the ability to resolve these strains into pathovars was 

enhanced, even to the extent that these differences were correlated 

with strains of a given host range. 



LITERATURE CITED 



1. Alvarez, A. M. , Benedict, A. A., and Mizumoto, C. Y. 1985. 
Identification of xanthomonads and grouping of strains of 
Xanthomonas campestris pv. campestris with monoclonal antibodies. 
Phytopathology 75:722-728. 

2. Barker, D., Schafer, M. , and White, R. 1984. Restriction sites 
containing CpG show a higher frequency of polymorphism in human 
DNA. Cell 36:131-138. 

3. Bender, C. L., and Cooksey, D. A. 1986. Indigenous plasmids in 
Pseudononas syringae pv. tonato : conjugative transfer and role in 
copper resistance. J. Bacteriol. 165:534-541. 

4. Beynon, J. L., Beringer, J. E., and Johnston, A. W. B. 1980. 
Plasmids and host-range in Rhizdjium leguminosarum and Rhizobium 
p^aseoli . J. Gen. Microbiol. 120:421-429. 

5. Boucher, C, Martinel, A., Barberis, P., Alloing, G., and 
Zischek, C. 1986. Virulence genes are carried by a megaplaanid of 
the plant pathogen Pseudonon as solanacearum. Mol. Gen. Genet 
205:270-275. 

6. Bradbury, J. F. 1984. Xanthononas Dowson 1939. pp. 199-210 in: 
Bergy's Manual of Systematic Bacteriology, Vol. 1. "n.r Krieg, and 
J.G Holt, eds. Williams and Wilkins, Baltimore. 964 pp. 

7. Brinkerhoff, L. A. 1963. Variability of Xanthomonas malvacearum : 
the cotton bacterial blight pathogen. Oklahoma State University 
Agricultural Experiment Station Bulletin T-98. 95 pp. 

8. Brinkerhoff, L. A. 1970. Variation in Xanthomonas malvacearum and 
Its relation to control. Annu. Rev. Phytopathol. 8:85-110. 

9. Brinkerhoff, L. A., and Presley, J. T. 1967, Effect of four day 
and night temperature regimes on bacterial blight reactions of 
immune, resistant, and susceptible strains of upland cotton. 
Phytopathology 57:47-51. 

10. Buchholz, W. G., and Thonashow, M. F. 1984. Conparison of T-DNA 
complements of Agrobacterium tumefaciens tumor-inducing plasmids 
with limited and wide host ranges. J. Bacteriol. 160:319-326. 



97 



98 



11. Bufton, L., Bruns, G. A. P., Magenis, R. E., Tomar, D. , Shaw, D., 
Brook, D., and Litt, M. 1986. Four restriction fracraent length 
polymorphisms revealed by probes from a single cosmid map to 
chromoscsne 19. Am. J. Hum. Genet. 38:447-460. 

12. Chase, A. R. 1986. Conparisons of three bacterial leaf spots of 
Hibiscus rosa-sinensis . Plant Dis. 70:334-336. 

13. Civerolo, E. L. 1985. Indigenous plasmids in Xanthomonas 
campestris pv. citri . Phytopathology 75:524-528. " ~ 

14. Close, T. J., Zaitlin, D., and Kado, C. I. 1984. Design and 
developnent of amplifiable broad-host-range cloning vectors: 
analysis of the vir region of Agrobacteri um tumefaciens plasmid 
pTiC58. Plasmid 12:111-118. 

15. Curiale, M. S., and Mills, D. 1983. Molecular relatedness among 
cryptic plasmids in Pseudomonas syringae pv. glycinea 
Phytopathology 73:1440-1444. '^ — — * 



16, 



Dahlbeck, D., Pring, D. R. , and Stall, R. E. 1977. Detection of 
covalently closed circular DNAs from Xanthomonas vesicatoria. 
(Abstr.). Proc. Amer. Phytopathol. Soc. 4:176. 



17. Dahlbeck, D., and Stall, R. e. 1979. Mutations for change of race 
m^ cultures of Xanthomonas vesicatoria . Phytopathology 

18. Daniels, M. J., Barber, C. E., Turner, P. C, Sawczyc, M. K., 
Byrde, R. j. w. , and Fielding, A. H. 1985. Cloning of genes 
involved in pathogenicity of Xanthomonas campestris pv. 
campestris using the broad host range cosmid pLAFRl. EMBO J 
13:3323-3328. 

19. Daub, M. E., and Hagedorn, D. J. 1981. Epiphytic populations of 
Pseudaxionas syringae on susceptible and resistant bean lines 
Phytopathology 71:547-550. 

20. Day, P. R. 1974. Genetics of Host-Parasite Interaction. W. H. 
Freeman and Co., San Francisco, CA. 238 pp. 

21. De Vos, P., Goor, M. , Gillis, M. , and De Ley, J. 1985. Riboscmal 
ribonucleic acid cistron similarities of phytopathogenic 
Pseudomonas species. J. Sys. Bacterid. 35:169-184. 

22. Ditta, G., Stanfield, S., Corbin, D., and Helinski, D. R. 1980. 
Broad host range DNA cloning system for Gram-negative bacteria: 
Construction of a gene bank of Rhizobium meliloti. Proc. Natl 
Acad. Sci. USA 77:7347-7351. 

23. Dye, D. W. 1958. Host specificity in Xanthomonas. Nature 
182:1813-1814. 



99 



24. Dye, D. PJ. , Bradbury, J. F., Goto, M. , Hayward, A. C. , Lelliott 
R. A., and Schroth, M. N. 1980. International standards for 
naming patnovars of phytopathogenic bacteria and a list of 
pathovar names and pathotype strains. Rev. Plant Pathol 
59:153-168. 

25. Dylan, T., lelpi, L., Stanfield, S., Kashyap, L., Douglas, C, 
Yanofsky, M. , Nester, E., Helinski, D, R., and Ditta, G. 1986 
Rhizobium meliloti genes required for nodule development are 
related to chromosomal virulence genes in Agrobacterium 
tumefaciens . Proc. Natl. Acad. Sci, USA 83:4403-440?; ' 

Ercolani, G. L. 1978. Pseudcmonas savastanoi and other bacteria 
colonizing the surface of olive leaves in the field. J. Gen 
Microbiol. 109:245-257. 



26. 



27. Essenberg, M. , Cason, E. T., Jr., Hamilton, B. , Brinkerhoff, L. 
A., and Richardson, R. K. 1979. Single-colonies of Xanthomonas 
malvacearum m susceptible and iirmune cotton leaves and the local 
resistance response to colonies in inmune leaves. Physiol. Plant 
Pathol. 15:53-68. ^ 

28. Farrar, W. E. , Jr. 1983. Molecular analysis of plasmids in 
epidemiologic investigation. J. Infect. Dis. 148:1-6. 

29. Follm, J.-C. 1983. Races of Xanthomonas campestr is pv 
malvacearum (Snith) Dye in western and cenral Africa. Cot Fib' 
Trop. 38:277-279. 

30. Fujimoto, D. K., and Vidaver, A. K. 1985. Analysis of strain 
variation m Xanthomonas campestris pv. phaseoli . (Abstr.) Proc. 
Sixth Int. Conf. Plant Pathogenic Bact., College Park, MD. 

^~' ?Qo^^^"^' ^' ^'' ■^^^^' ^°"^ Plaa^id DNA variants distinguished in 
^^.^.f!.^"'^^^^ citrus canker epiphytotic. (Abstr.) Phytopathology 

32. Gabriel, D. W. 1986. Specificity and gene function in 
plant-pathogen interactions. ASM Nev;s 52:19-25. 

33. Gabriel, D. W. , Burges, A., and Lazo, G. R. 1986. Gene-for-gene 
interactions of five cloned avirulence qenes fron Xanthc monas 
campestris pv. malvacearum with specific resistance genes in 
cotton. Proc. Natl. Acad. Sci. USA. 83:6415-6419. 

34. Gabriel, D. W. , Burges, A. R., Lazo, G. R. , and Roffey, R. 1986. 
Xanthcmonas campestris pvs. citri, alfalfae , and phaseoli are 
^^"f^^^^l^y ^^ pathologically related. (Abstr.). Phytopathology 

/ D • _LU / u • 



100 



35. Gallie, D. R., Zaitlin, D., Perry, K. L., and Kado, C. I. 1984 
Characterization of the replication and stability regions of 
Agrooacterium tumefaciens plasmid pTAR. J. Bacteriol 
157:739-745. " ^«riox. 

36. Gerhardt, P., Murray, R. E. E., Costilow, R. N., Nester, E. W. , 
Wood, W. A., Krieg, M. R., and Phillips, G. B. 1981. Manual of 
Metnods for General Bacteriology. American Society for 
Microbiology, Washington, DC. 524 pp. 

37. Goto M., and Starr, M. P. 1972. Phage-host relationships of 
Xanthononas citri conpared with those of other Xanthomonas. Ann 
Phytopatol. Soc. Japan. 38:226-248. 

38. Gottlieb, P., and Rudner, R. 1985. Restriction site polymorphism 
of ribosomal ribonucleic acid gene sets in members of the genus 
Bacillus . Int. J. Sys. Bacteriol. 35:244-252. 

39. Haas, J. M. , Fett, W. F., and Flerring, D. J. 1985. Detection and 
initial characterization of plasmids in Xanthomon as camoestris 
pv. glycines (Abstr.). Proc. Sixth Int. Conf. Plant Pathogenic 
Bact., College Park, MD. 

40. Hartung, j. s., and Civerolo, E. L. 1986. Genomic fingerprints of 
Xanthomonas campestris pv. citri strains from Asia, South 
America, and Florida. (Abstr.). Phytopathology 76:1137. 

41. Helentjaris, T. , King, G., Slocum, M. , Siedenstang, C., and 
Wegman, S. 1985. Restriction fragment polymorphisms as probes for 
plant diversity and their development as tools for applied plant 
breeding, plant Mol. Biol. 5:109-118. 

42. Heumann, W. , Rosch, A., Springer, R., Wagner, E. , and Winkler, K. 
P. 1984. In rhizobiaceae five different species are produced by 
rearrangements of one genome, induced by DNA-damaging agents. 
Mol. Gen. Genet. 197:425-436. ^ y y 

43. Hoekema, A., De Pater, B. S., Felinger, A. J., Hooykaas, P. J. 
J., and Schilperoort, R. A. 1984. The limited host ranae of an 
Agrobacteriurr. tumefaciens strain extended by a cytokinin gene 
^rom a wide host range T-region. EMBO J. 3:3043-304?". 

44. Howie, W.,^ and Suslow, T. 1986. Effect of antifungal compound 
biosynthesis on cotton root colonization and Pytliium" " suppression 
by a strain of Pseudcmonas fluorescens and its antifungal minus 
isogenic mutant. (Abstr.). Phytopathology 76:1069. 

45. Hunter, R. E., Brinkerhoff, L. A., and Bird, L. S. 1968 The 
development of a set of upland cotton lines for differentiating 
races of Xanthomonas malvacearum . Phytopathology 58:830-832. 



101 



46. Kado, C. I., and Liu, S. T. 1981. Rapid procedure for detection 
and isolation of large and small plasmids. J. Bacteriol 
145:1365-1373. a<^c.erioi. 

47. Kamper, S. M. , French, W. J., and deKloet, S. R. 1985. Genetic 
relationships of some fastidious xylem- limited bacteria. Int. j 
Sys. Bacteriol. 35:185-188. 

48. Kao, J. C, Perry, K. L., and Kado, C. I. 1982. Indoleacetic acid 
canplementation and its relation to host range specifving genes 
on the Ti plasmid of Agrobacterium tumefaciens. Mol. Gen. Genet 
188:425-432. ^ne.. 



m ; 



49. Klement, Z. 1983. Hypersensitivity, pp. 150-177 ... 
Phytopathogenic Prokaryotes, Vol. 2. M. S. Mount, and G. H. Lacy, 
eds. Academic Press, New York, NY. 506 pp. 

50. Lacy, G. H., and Patil, S. S. 1982. Why genetics?, pp. 221-228 
m: Phytopathogenic Prokaryotes, Vol 2. M. S. Mount, and G. H. 
Lacy, eds. Academic Press, Inc., New York, NY. 506 pp. 

51. Laemmli, U. K. 1970. Cleavage of structural proteins during the 
assembly of the head of bacteriophage T4. Nature 227:680-685. 

52. Lazo, G. R. , and Gabriel, D. W. 1984. Described 'races' of 
Xanthomonas campestris pv. malvacearum are mixtures. (Abstr ) 
Phytopathology 74:837. 

53. Lazo, G. R., and Gabriel, D. W. 1985. Use of plasmid DNAs to 
aifferentiate pathovars of Xanthomonas campestris. (Abstr ) 
Phytopathology 75:1320. " 

54. Lazo, G. R. , and Gabriel, D. W. Conservation of plasmid DNA 
sequences and pathovar identification of strains of Xanthomonas 
campestris . Phytopathology (in press) . " 

55. Lazo, G. R., Gabriel, D. W. , and Roffey, R. 1986. Differentiating 
parhovars of Xanthomonas campestris without pathogenicity tests 

(Abstr.). Phytopathology 76:1076. 

56. Lelliott, R. A., and Dickey, R. s. 1984. Erwinia Winslow, 
Broadhurst, Buchanan, Krumwiede, Rogers and Smith 1920. pp. 
469-476 in: Bergy's Manual of Systematic Bacteriology, Vol l" 
N. R. Krieg, and J. G. Holt, eds. Williams and Wilkins' 
Baltimore, MD. 964 pp. 

57. Leyns, F., De Cleene, M. , Swing, j., and De Ley, J. 1984. The 
host range of the genus Xanthcmonas . Bot. Rev. 50:308-356. 

58. Li, W.-H. 1986. Evolutionary change of restriction cleavage sites 
and phylogenetic inference. Genetics 113:187-213. 



59. Liew, K W., and Alvarez, A. M. 1981. Phage typing and lysotyoe 
71 -274-275^'°" ""^ Xanthomonas campestris . Phytopathology 

60. Lin B., and Chen, S. 1978. Multi-plasmid in Xanthom onas 
;^"^°t"^ (ADstr.). Third Int. Congress Plant Patho l., Munchen, 
Germany. ' 



102 



61, 



Litt, M., and White, R. L. 1985. A highly polymorphic locus in 
himan DNA revealed by cosmid-derived probes. Proc. Natl Acad 
Sci. USA 82:6206-6210. 



62. Long, S. R. , Buikema, W. , and Ausebel, F. M. 1982. Cloning of 
Rhizobium meliloti nodulation genes by direct complementation of 
Nod mutants. Nature 298:485-488. 

63. Loper, J. E. 1986. Role of fluorescent siderophore production in 
biocontrol of Pythium ultimum by Pseudomonas f lucre scens strain 
3551. (Abstr.). Phytopathology 76:1069^; ' 

64. Loper, J. E., and Kado, C. I. 1979. Host range conferred by the 
virulence- specifying plasmid of Agrobacte rium tumefaciens J 
Bacteriol. 139:591-596. " 

65. Maas,_J._L., Finney, M. M. , Civerolo, E. L., and Sasser, M. 1985 
Association of an unusual strain of Xanthomonas c^pestris with 
apple. Phytopathology 75:438-445. 

66. Maniatis, T., Fritsch, E. F., and SairfDrook, J. 1982. Molecular 
Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold 
Spring Harbor, N.Y. 545 pp. 

67. McClean, P. E., and Hanson, M. R. 1986. Mitochondrial DNA 
sequence divergence among Lycopersicon and related Solanum 
species. Genetics 112:649-667."^ ~ 

68. McMaster, G. K.,^ Samulski, R. j., stein, J. L., and Stein, G. S. 
1980. Rapid perification of covalently closed circular DNAs of 
10^4^-54^ ^^^^^'^^ ^"^ animal tumor viruses. Anal. Biochem. 

69. McNally, K., and Gabriel, D. W. 1984. Useful minimal culture 
media for Xanthomonas campestris pv. malvacearum. (Abstr.) 
Phytopathology 74:875. 

70. Minsavage, G. V., and Schaad, N. W. 1983. Characterization of 
memorane proteins of Xanthomonas campestris pv. r^ampestris 
Phytopathology 73:747-755^ " ' * 

71. Mulrean, E. N., and Schroth, M. N. 1982. Ecology of Xanthcmonas 
campestris pv. juglandis on Persian (English) walnuts. 
Phytopathology 72:434-438. 



103 



72. Murai, N., Skoog, F., Doyle, M. E., and Hanson, R. S. 1980. 
Relationships between cytokinin production, presence of plasmids, 
and fasciation caused by strains of Corynebacterium fascians . 
Proc. Natl. Acad. Sci. USA 77:619-623. 

73. Murata, N. , and Starr, M. P. 1973. A concept of the genus 
Xanthomonas and its species in the light of segmental hanology of 
deoxyribonucleic acids. Phytopath. Z. 77:285-323. 



74, 



Nei, M., and Li, W.-H. 1979. Mathematical model for studying 
genetic variation in terms of restriction endonucleases. Proc. 
Natl. Acad. Sci. USA 76:5269-5273. 



75. Palleroni, N. J, 1984. Family I. Pseud omonadaceae Winslow, 
Broadhurst, Buchanan, Krumwiede, Roger and Smith 1957. pp. 
140-199 in: Bergy's Manual of Systematic Bacteriology, Vol 1. N. 
R. Krieg, and J. G. Holt, eds. Williams and Wilkins, Baltimore, 
MD. 

76. Panopoulos, N. J., and Peet, R. C. 1985. The molecular genetics 
of plant pathogenic bacteria and their plasmids. Annu. Rev. 
Phytopathol. 23:381-419. 

77. Piwowarski, J., and Shaw, P. D. 1982. Characterization of 
plasmids from plant pathogenic pseudomonads. Plasmid 7:85-94. 

78. Poplawsky, A. R. , Peng, Y. F., and Ellingboe, A. H. 1986. 
Bacterial Tn5 mutants affected in antibiosis to Gaeumanmcaiiyces 
graminis var. tritici . (Abstr.). Phytopathology 76:1069. 

79. Quant, R. L., and Mills, D. 1984. An integrative plasmid and 
multiple-sized plasmids of Pseudcmonas syringae pv. phaseolicola 
have extensive homology. Mol. Gen. Genet. 193:459-4667 

80. Randhawa, P. S., and Civerolo, E. L. 1985. Plasmids in 
XanthCTnonas campestris pv. pruni (Abstr.). Proc. Sixth Int. Conf. 
Plant Pathogenic Bact. , College Park, MD. 

81. Sasser, M. 1983. Fatty acid analysis of the genus Xanthononas . 
in: First Fallen Leaf Lake Conf. On the Genus Xanthcmonas . Fallen 
Leaf Lake, Q\. 

82. Schaad, N. W. 1980. Laboratory Guide for Identification of Plant 
Pathogenic Bacteria. American Phytopathological Society, St. 
Paul, MN. 72 pp. 

83. Schleifer,^ K. H., Ludwig, W. , Kraus, J., and Festl, H. 1985. 
Cloned ribosomal ribonucleic acid genes from Pseudomonas 
aeruginosa as probes for conserved deoxyribonucleic acid 
sequences. Int. J. Sys. Bacterid. 35:231-236. 



104 



84. Schnathorst, W. C. 1966. Unaltered specificity in several 
xanthcsnonads after repeated passage through Phaseolus vulgaris. 
Phytopathology 56:58-60. 

85. Schroth, M. N., and Hildebrand, D. C. 1983. Toward a sensible 
taxonomy of bacterial plant pathogens. Plant Dis. 67:128. 

86. Schroth, M. N., Hildebrand, D. C. , and Vitanza, V. 1970. 
Pathogenic variation and overlapping host ranges in Pseudononas 
phaseolicola , P. glycinea , and P. mori . (Abstr.). Phytopathology 
60:1313. 

87. Silhavy, T. J., Berman, M. L. , and Enquist, L. W. 1984. 
Experiments with Gene Fusions. Cold Spring Harbor Laboratory, 
Cold Spring Harbor, NY. 

88. Stall, R. E., Loschke, D. C, and Jones, J. B. 1986. Linkage of 
copper resistance and avirulence loci on a self-transmissible 
plasmid in Xanthcmonas campes tris pv. vesicatoria. Phytopatholoay 
76:240-243. 

89. Starr, M. P. 1983. The genus Xanthcmonas . pp. 742-763 in: The 
Prokaryotes, Vol. 2. M. P. Starr, H. Stolp, H, G. Truper, A. 
Balov/s, and H. G. Schlegel, eds. Springer-Verlag, New York, NY. 

90. Surico, G., Canai, L., and Kosuge, T. 1984. Pathogenicity of 
strains of Pseudomonas syringae pv. savastanoi and their 
indoleacetic acid-deficient mutants on olive and oleander. 
Phytopathology 74:490-493. 

91. Sytsma, K. J., and Gottlieb, L. D. 1986. Chloroplast IMh evidence 
for the origin of the genus Heterogaura from a species of Clarkia 
(Onagraceae) . Proc. Natl. Acad. Sci. USA 83:5554-5557. 

92. Szabo, L. J., and Mills, D. 1984. Characterization of eight 
excision plasmids of Pseudononas syringae pv. phaseoli cola. Mol. 
Gen. Genet. 195:90-95. 

93. Tait, R. C, Close, T., Lundquist, R., Hagiya, M. , Rodriguez, R., 
and Kado, C. I. 1983. Construction and characterization of a 
versatile broad host range DNA cloning system for gram-negative 
bacteria. Biotechnology 1:269-275. 

94. Thaveechai, N., and Schaad, N. W. 1984. Comparison of different 
immunogen preparations for serological identification of 
Xanthomonas campestris pv. campestris. Phytopathology 
74:1065-1070. . 

95. Thaveechai, _ N. , and Schaad, N. W. 1986. Immunochemical 
characterization of a subspecies-specific antigenic determinant 
of a membrane protein extract of Xanthcmonas campestris pv. 
campestris . Phytopathology 76:148-153. 



105 



96. 



Thayer, D. W. , Lowther, S. V., and Phillips, J. G. 1984 
Cellulolytic activities of strains of the genus Cellulomonas" 
Int. J. Sys. Bacteriol. 34:432-438. 



97. Watson, B. , Currier, T. C, Gordon, M. P., Chilton, M.-D., and 
Nester, E. W. 1975. Plasmid required for virulence of 
Agrobacterium tumefaciens . J. Bacteriol. 123:255-264. 

98. Young, J. M. , Dye, D. W. , Bradbury, J. F., Panagopoulos, C. G., 
and Robbs, C. F. 1978. A proposed nomenclature and classification 
for plant pathogenic bacteria. N. Z. J. Agric. Res. 21:153-177. 

99. Yuen, G. Y. K., and Alvarez, A. M. 1985. Aberrant simptoras on 
cabbage caused by strains of Xanthononas camoestris. (Abstr ) 
Phytopathology 73:1382. " 

100. Yuen, G. Y., and Schroth, M. N. 1986. Interactions of Pseudanonas 
fluorescens strain E6 with ornamental plants and its effect on 
the composition of root-colonizing microflora. Phytopathology 
76:176-180. j t^ ^y 



BIOGRAPHICAL SKETCH 

Gerard Raymond Lazo was born in Dallas, Texas on August 25, 1957, 
to Francisco Gerardo and Suzanne Marie Lazo. He completed his primary 
and secondary education in Weslaco, Texas, where he graduated from 
Weslaco High School in 1975. He received the degree of Bachelor of 
Science in microbiology in 1979 from Texas A&M University. He 
continued his education under the supervision of Luther S. Bird 
conducting research on fungal and bacterial diseases of cotton and 
received a Master of Science degree in plant pathology frcm Texas A&M 
University in 1984. Gerard was married to Maria Alicia Gonzales in 
1983 while attending Texas A&M University. In the fall of 1983 he 
began to pursue the degree of Doctor of Philosophy in plant pathology 
at Oklahcsna State University and later continued his education at the 
University of Florida under the supervision of Dean W. Gabriel. Upon 
ccmpleting his degree requirements he plans to serve as a postdoctoral 
research associate with Rcbert A. Ludwig at Thimann Laboratories, 
University of California, Santa Cruz. Gerard's permanent address is 
1807 Briarcrest Lane, Arlington, Texas 76012. 



106 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 



I. 



•'-! ,../'>x!^^ 



/J 



^f) 



Dean W. Gabriel) Chairman 
Assistant Professor of Plant 
Pathology 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 




Robert E. Stall 

Professor of Plant Pathology 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 




Daryl R. Prj 

Professor of Plant Pathology 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is 
fully adequate, in scope and quality, as a dissertation for the degree 
of Doctor of Philosophy. 



r 



lr\ ojJi ^ :12 



a/y) 



Mark J. Basset 
Professor of H&rticultural 
Science 



This dissertation was submitted to the Graduate Faculty of the College 
of Agriculture and to the Graduate School and was accepted as partial 
fulfillment of the requirements for the degree of Doctor of 
Philosophy. 



May, 1987 




a^^ ^. Jy^ 



Dean ,/3Sol lege of Agriculture 



Dean, Graduate School